HANDBOOK OF FISH BIOLOGY AND FISHERIES Volume 1 Also available from Blackwell Publishing: Handbook of Fish Biology and Fisheries Edited by Paul J.B. Hart and John D. Reynolds Volume 2 Fisheries Handbook of Fish Biology and Fisheries VOLUME 1 FISH BIOLOGY

EDITED BY Paul J.B. Hart Department of Biology University of Leicester

AND

John D. Reynolds School of Biological Sciences University of East Anglia © 2002 by Blackwell Science Ltd a Blackwell Publishing company

Chapter 8 © British Crown copyright, 1999

BLACKWELL PUBLISHING 350 Main Street, Malden, MA 02148‐5020, USA 108 Cowley Road, Oxford OX4 1JF, UK 550 Swanston Street, Carlton, Victoria 3053, Australia

The right of Paul J.B. Hart and John D. Reynolds to be identified as the Authors of the Editorial Material in this Work has been asserted in accordance with the UK Copyright, Designs, and Patents Act 1988.

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs, and Patents Act 1988, without the prior permission of the publisher.

First published 2002 Reprinted 2004

Library of Congress Cataloging‐in‐Publication Data has been applied for.

Volume 1 ISBN 0‐632‐05412‐3 (hbk) Volume 2 ISBN 0‐632‐06482‐X (hbk) 2‐volume set ISBN 0‐632‐06483‐8

A catalogue record for this title is available from the British Library.

Set in 9/11.5 pt Trump Mediaeval by SNP Best‐set Typesetter Ltd, Hong Kong Printed and bound in the United Kingdom by TJ International Ltd, Padstow, Cornwall.

The publisher’s policy is to use permanent paper from mills that operate a sustainable forestry policy, and which has been manufactured from pulp processed using acid‐free and elementary chlorine‐free practices. Furthermore, the publisher ensures that the text paper and cover board used have met acceptable environmental accreditation standards.

For further information on Blackwell Publishing, visit our website: http://www.blackwellpublishing.com Contents

List of Contributors, x Preface, xii List of Abbreviations, xiv

1 BANISHING IGNORANCE: UNDERPINNING FISHERIES WITH BASIC BIOLOGY, 1 Paul J.B. Hart and John D. Reynolds 1.1 Introduction, 1 1.2 Global fisheries, 1 1.3 The quest for knowledge, 3 1.4 Part 1: Biodiversity, 4 1.5 Part 2: Production and population structure, 5 1.6 Part 3: Fish as predators and prey, 7 1.7 Part 4: Fish in ecosystems, 8 1.8 Ignorance banished?, 9 1.9 Conclusions, 10 Part 1: Biodiversity, 13

2 PHYLOGENY AND SYSTEMATICS OF FISHES, 15 A.C. Gill and R.D. Mooi 2.1 Introduction, 15 2.2 Phylogenetic methods and classification, 15 2.3 Fish diversity and phylogeny, 20 2.4 Conclusions, 36

3 HISTORICAL BIOGEOGRAPHY OF FISHES, 43 R.D. Mooi and A.C. Gill 3.1 Introduction, 43 3.2 Concepts and methods, 44 3.3 Distribution, faunal composition and historical biogeography by region, 47 3.4 Conclusions, 62 vi Contents

Part 2: Production and Population Structure, 69

4 THE PHYSIOLOGY OF LIVING IN WATER, 71

Ole Brix

4.1 Introduction, 71 4.2 Buoyancy, or coping with pressure, 72 4.3 Swimming, 75 4.4 Osmoregulatory problems in fresh and salt water, 78 4.5 Respiration and special adaptations for living in low oxygen, 82 4.6 Digestion and absorption, 90 4.7 Bioluminescence, 91 4.8 Conclusions, 92

5 ENVIRONMENTAL FACTORS AND RATES OF DEVELOPMENT AND GROWTH, 97

Malcolm Jobling

5.1 Introduction, 97 5.2 Terminology of life-history stages, 97 5.3 Development and growth during early life history, 99 5.4 Growth models and equations, 102 5.5 Age determination, back-calculation and validation techniques, 104 5.6 Length–weight relationships and indices of condition and growth, 107 5.7 Energy budget and bioenergetics: energy partitioning and storage, 109 5.8 Growth at different latitudes: models of growth compensation, 113 5.9 Estimating food consumption, 115 5.10 Conclusions, 117

6 RECRUITMENT: UNDERSTANDING DENSITY-DEPENDENCE IN FISH POPULATIONS, 123

Ransom A. Myers

6.1 Introduction, 123 6.2 The link between spawner abundance and subsequent recruitment, 124 6.3 Generalities through meta-analysis, 129 6.4 Carrying capacity, 130 6.5 Variability in recruitment, 131 6.6 At what life-history stage does density-dependent mortality occur?, 131 6.7 Estimating density-dependent mortality from long-term surveys, 133 6.8 Pelagic egg, larval and juvenile stages, 136 6.9 Future research, 141 6.10 Conclusions, 144 Contents vii

7 LIFE HISTORIES OF FISH, 149

J.A. Hutchings

7.1 Introduction, 149 7.2 Influence of survival and growth rate on age, size and reproductive effort at maturity, 152 7.3 Offspring size and number strategies ,158 7.4 Alternative life-history strategies, 162 7.5 Effects of fishing on life history, 165 7.6 Conclusions, 167

8 MIGRATION, 175

Julian Metcalfe, Geoff Arnold and Robert McDowall

8.1 Introduction, 175 8.2 Exploitation and ecology, 178 8.3 Fish migrations, 179 8.4 Migratory mechanisms, 189 8.5 Techniques, 191 8.6 Distribution and genetics, 192 8.7 Fishery applications, 194 8.8 Conclusions, 194

9 GENETICS OF FISH POPULATIONS, 200

Robert D. Ward

9.1 Introduction, 200 9.2 Genetic tools, 200 9.3 Statistical tools, 205 9.4 Specimen and species identification, 206 9.5 Fish population genetics, 207 9.6 Genetics of sex determination in fish, 218 9.7 Conclusions, 218

10 BEHAVIOURAL ECOLOGY OF REPRODUCTION IN FISH, 225

Elisabet Forsgren, John D. Reynolds and Anders Berglund

10.1 General introduction, 225 10.2 Introduction to breeding systems, 225 10.3 Parental care, 228 10.4 Sexual selection, 230 10.5 Mating patterns, 236 10.6 Reproductive behaviour and life histories, 238 10.7 Reproductive behaviour and exploitation, 239 10.8 Conclusions, 241 viii Contents

Part 3: Fish as Predators and Prey, 249

11 FISH FORAGING AND HABITAT CHOICE: A THEORETICAL PERSPECTIVE, 251

Gary G. Mittelbach

11.1 Introduction, 251 11.2 Foraging behaviour and diet choice, 252 11.3 Foraging models and fish growth, 254 11.4 Feeding rate and group size, 255 11.5 Foraging and habitat selection, 256 11.6 Conclusions, 262

12 FEEDING ECOLOGY OF PISCIVOROUS FISHES, 267

Francis Juanes, Jeffrey A. Buckel and Frederick S. Scharf

12.1 Introduction, 267 12.2 Adaptations for piscivory, 267 12.3 Components of predation, 271 12.4 Prey type and size selectivity, 274 12.5 Predator-size and prey-size relationships, 275 12.6 Population regulation, 277 12.7 Methods of studying predation in the field, 278 12.8 Implications for conservation and management, 279 12.9 Conclusions, 279

13 FISH AS PREY, 284

J. Krause, E.M.A. Hensor and G.D. Ruxton

13.1 Introduction, 284 13.2 Immobility, 285 13.3 Mobility, 287 13.4 Conclusions, 293

Part 4: Fish in Ecosystems, 299

14 TROPHIC ECOLOGY AND THE STRUCTURE OF MARINE FOOD WEBS, 301

Nicholas V.C. Polunin and J.K. Pinnegar

14.1 Introduction, 301 14.2 Food chains and food webs, 302 14.3 Interaction strength in food webs, 312 14.4 Implications of food webs and trophodynamics for fish and fisheries science, 314 14.5 Conclusions, 316 Contents ix

15 COMMUNITY ECOLOGY OF FRESHWATER FISHES, 321 Lennart Persson 15.1 Introduction, 321 15.2 Community patterns and basic ecological performance, 322 15.3 Competition and predation as structuring forces, 325 15.4 Fish community structure, productivity and habitat structure, 328 15.5 Effects of fish on lower trophic components, 331 15.6 From individual-level processes to population dynamics, 334 15.7 Conclusions, 337

16 COMPARATIVE ECOLOGY OF MARINE FISH COMMUNITIES, 341 K. Martha M. Jones, Dean G. Fitzgerald and Peter F. Sale 16.1 Introduction, 341 16.2 Biodiversity, 342 16.3 Habitat associations, 346 16.4 Differences in tropical and temperate production cycles, 350 16.5 Variation in recruitment dynamics, 351 16.6 Conclusions, 352

17 INTERACTIONS BETWEEN FISH, PARASITES AND DISEASE, 359 I. Barber and R. Poulin 17.1 Introduction, 359 17.2 Fish parasite diversity, 360 17.3 Evolution of host–parasite relationships, 365 17.4 Effects of parasites on fish population ecology, 374 17.5 Socioeconomic and human health implications of fish parasites, 376 17.6 Controlling parasite infections, 382 17.7 Recent applications of fish parasitology, 383 17.8 Conclusions, 384

Index, 390 Contributors

Geoff Arnold The Centre for Environment, Elisabet Forsgren Department of Marine Fisheries and Aquaculture Science, Lowestoft Ecology, Göteborg University, Kristineberg Ma- Laboratory, Pakefield Road, Lowestoft, Suffolk rine Research Station, SE-450 34 Fiskebäckskil, NR33 0HT, UK. Sweden [email protected] [email protected]

The Natural History Museum, I. Barber Edward Llwyd Building, Institute of A.C. Gill Cromwell Road, London SW7 5BD, UK Biological Sciences, The University of Wales [email protected] Aberystwyth, Aberystwyth, SY23 3DA, UK. [email protected] Paul J.B. Hart Department of Biology, University of Leicester, Leicester LE1 7RH, UK Anders Berglund Animal Ecology, Evolution- [email protected] ary Biology Centre, Uppsala University, Norbyvä- gen 18 D, SE-752 36 Uppsala, Sweden E.M.A. Hensor School of Biology, University [email protected] of Leeds, Leeds LS2 9JT, UK [email protected] Ole Brix Department of Zoology, University of Bergen, Allégaten 41, N-5007 Bergen, Norway J.A. Hutchings Department of Biology, [email protected] Dalhousie University, Halifax, NS B3H 4J1, Canada [email protected] Jeffrey A. Buckel Department of Zoology, Center for Marine Sciences and Technology, North Malcolm Jobling The Norwegian College of Carolina State University, 303 College Circle, Fishery Science, University of Tromsø, N-9037 Morehead City, NC 28557, USA Tromsø, Norway [email protected] [email protected]

Dean G. Fitzgerald Department of Natural K. Martha M. Jones Institute of Marine and Resources, Biological Field Station, Cornell Uni- Coastal Sciences, Rutgers University Marine versity, 900 Shackelton Point Road, Bridgeport, Field Station, 800 c/o 132 Great Bay Boulevard, NY 13030, USA Tuckerton, NJ 08087, USA [email protected] [email protected] Contributors xi

Francis Juanes Department of Natural J.K. Pinnegar The Centre for Environment, Resources Conservation, University of Fisheries and Aquaculture Science, Lowestoft Massachusetts, Amherst, MA 01003-4210, USA Laboratory, Pakefield Road, Lowestoft, Suffolk [email protected] NR33 0HT, UK [email protected] J. Krause School of Biology, University of Leeds, Leeds LS2 9JT, UK Nicholas V.C. Polunin Department of Marine [email protected] Sciences and Coastal Management, University of Newcastle, Newcastle upon Tyne, NE1 7RU, UK Robert McDowall National Institute of [email protected] Water and Atmospheric Research, PO Box 8602, Christchurch, New Zealand R. Poulin Department of Zoology, University [email protected] of Otago, PO Box 56, Dunedin, New Zealand [email protected] Julian Metcalfe The Centre for Environment, School of Biological Fisheries and Aquaculture Science, Lowestoft John D. Reynolds Sciences, University of East Anglia, Norwich NR4 Laboratory, Pakefield Road, Lowestoft, Suffolk 7TJ, UK NR33 0HT, UK [email protected] [email protected] G.D. Ruxton Division of Environmental and Kellogg Biological Gary G. Mittelbach Evolutionary Biology, Graham Kerr Building, Station, Michigan State University, Hickory Glasgow University, Glasgow G12 8QQ, UK Corners, MI 49060, USA [email protected] [email protected] Peter F. Sale Department of Biology and Great R.D. Mooi Milwaukee Public Museum, 800 Lakes Institute for Environmental Research, Uni- West Wells Street, Milwaukee, WI 53233-1478, versity of Windsor, Windsor, Ontario N9B 3P4, USA Canada [email protected] [email protected]

Ransom A. Myers Killam Chair of Ocean Frederick S. Scharf Northeast Fisheries Sci- Studies, Department of Biology, Dalhousie Uni- ence Center, James J. Howard Marine Sciences versity, Halifax, NS B3H 4J1, Canada Laboratory, 74 Magruder Road, Sandy Hook, [email protected] Highlands, NJ 07732, USA [email protected] Lennart Persson Department of Ecology and Environmental Science, Animal Ecology, Umeå Robert D. Ward CSIRO Research, PO Box University, SE-901 87 Umeå, Sweden 1538, Hobart, Tasmania 7001, Australia [email protected] [email protected] Preface

The goal of the two volumes of the Handbook of coverage of fish biology simply because the topics Fish Biology and Fisheries is to help integrate the are interesting in their own right, and because study of fish biology with the study of fisheries. the borders between pure and applied research are One might not expect these two subjects to need fuzzy. We therefore decided not to restrict infor- further integration. However, strong declines in mation according to its direct relevance to fish- many fish stocks around the globe, combined with eries as this subject is understood today. We hope growing concerns about the impact of fisheries the result will be of value to undergraduates and on marine and freshwater biodiversity, are raising graduates looking for information on a wide vari- new questions about aspects of fish biology that ety of topics in fisheries science. The books are also have traditionally dwelt outside mainstream aimed at researchers who need up-to-date reviews fisheries research. Thus, fisheries biologists and of topics that impinge on their research field but managers are increasingly asking about aspects may not be central to it. The information should of ecology, behaviour, evolution and biodiversity also be useful to managers and decision makers that had traditionally been studied by different who need to appreciate the scientific background people who attend different conferences and pub- to the resources they are trying to manage and lish in different journals. By bringing these people conserve. and their subjects together in the two volumes of In the first volume, subtitled Fish Biology, our this Handbook, we hope to foster a better two-way introductory chapter explores the underpinnings flow of information between the studies of fish of fisheries biology and management by basic biology and fisheries. research on fish biology. Part 1 then examines A tradition runs through the prefaces of the systematics and biogeography of fishes, including other volumes in this series whereby the editors methods for determining phylogenetic relation- distance themselves from a literal translation of ships and understanding spatial patterns of diver- the word ‘Handbook’. In keeping with this tradi- sity. Part 2 examines production and population tion, we wish to make clear that this is not a cook- structures of fishes, beginning with chapters on book of recipes for how to study and manage fish physiology and growth, followed by recruitment, populations. Instead, we have tried to produce a life histories, migration, population structure and pair of reference books that summarize what is reproductive ecology. Part 3 considers fishes as known about fish biology and ecology, much of predators and as prey, making use of conceptual which is relevant to assessment and management advances in behavioural ecology to link predator– of fish populations and ecosystems. Of course, prey interactions to the environment. In Part much of the material in the first volume may never 4 we scale up from individual interactions to find applications in fisheries, and that is fine with communities and ecosystems. The chapters us. We encouraged our authors to provide a wide include comparisons of freshwater and marine Preface xiii communities, as well as interactions between Sandford at Blackwell Science who has overseen it. fishes and parasites. The final stages of production have been greatly In the second volume, subtitled Fisheries, we helped by two people. Valery Rose of Longworth begin with a chapter that considers the human Editorial Services worked patiently and efficiently dimension of fisheries management. Part 1 then to shepherd the book through the copy editing and gives background information for fisheries, in- production stages, and Monica Trigg, did an excel- cluding fishing technology, marketing, history of lent job with the gargantuan task of constructing fisheries, and methods of collecting and present- the two indices. Paul Hart would like to thank the ing data. Part 2 provides fundamental methods of University of Bergen, Department of Fisheries and stock assessment, including surplus production Marine Biology, for providing him with space and models, virtual population analyses, methods for support during the final editing of the manuscript. forecasting, length-based assessments, individual- John Reynolds would like to thank his many col- based models and economics. We have also tried leagues at the University of East Anglia and at the to consolidate this information by reviewing the Centre for Environment, Fisheries, and Aquacul- various options available for modelling fisheries, ture Science (Lowestoft) who have helped build including the pros and cons of each. Part 3 explores bridges between pure and applied research. wider issues in fisheries biology and management, The editors and publishers are grateful to copy- including the use of marine protected areas, con- right holders for providing permission to repro- servation threats to fishes, ecosystem impacts and duce copyright material. If any sources have been recreational fishing. inadvertently overlooked, the publishers will We are very grateful to our 54 authors from 10 be pleased to rectify this omission at the first countries for their hard work and patience while opportunity. we attempted to herd them all in the same general direction. We also thank Susan Sternberg who first Paul J.B. Hart and John D. Reynolds suggested that we take on this project, and Delia Leicester and Norwich Abbreviations

AFLP amplified fragment length MSVPA multispecies virtual polymorphism population analysis CPUE catch per unit effort NTP nucleotide triphosphates CV coefficient of variation OFT optimal foraging theory EPIC exon-primed intron-crossing PCR PCR polymerase chain reaction ESD environmental sex determination RAPD randomly amplified FAO Food and Agriculture polymorphic DNA Organization RFLP restriction fragment length GFR glomerular filtration rate polymorphism GSI gonadosomatic index RQ respiratory quotient IBI indices of biotic integrity SCFA short-chain fatty acids ICES International Council for the SSVPA single-species virtual Exploration of the Seas population analysis ICLARM International Center for Living TMAO trimethylamine oxide Aquatic Resources Management VBGF von Bertalanffy growth function IFD ideal free distribution WCMC World Conservation JAM judicious averaging method Monitoring Centre MPA marine protected area WRI World Resources Institute MSA mixed stock analysis YOY young of the year 1 Banishing Ignorance: Underpinning Fisheries with Basic Biology

PAUL J.B. HART AND JOHN D. REYNOLDS

1.1 INTRODUCTION Bergen, covering perhaps 2km2, contains all the el- ements of the community that the two volumes of Tourists mingle with local people among the this book are designed to serve. In other parts of the stalls set out by Zachariasbryggen, Bergen, world the various institutions of fisheries are not Norway. Young men and women in bright orange as close to each other as they are in Bergen but the oilskin overalls, fluent in several languages, sell elements will be some local version of what exists fresh crabs (Cancer pagurus), lobsters (Homarus around Vågen. The heart of the system is the fisher gammarus), mackerel (Scomber scombrus), cod and the market at which the produce is landed and (Gadus morhua) and farmed salmon (Salmo salar) sold. to Japanese, German, British, American and Dutch tourists. People from Bergen buy their supper as they pass by during their lunch break. Unseen 1.2 GLOBAL FISHERIES by the mill of tourists and locals are the fishers who started from port at four in the morning to set According to the United Nations Food and their nets, go trawling or lift their crab pots. If you Agricultural Organization (FAO), there was a travel about a kilometre along the western shore of steady increase of fish catches until the middle of Vågen, the bay surrounded by the centre of Bergen, the 1990s when the catch began to level off (Fig. you come to Nordnes, the location of the Institute 1.1). Recent work by Watson and Pauly (2001) has for Marine Research, the public aquarium and the shown that in reality the total marine catch of fish offices of the Norwegian Fisheries Ministry. The has been declining by some 10% a year since 1988. tall building is full of biologists working on fish The apparent continued increase until the mid- stock assessments and on research into the marine 1990s was due to inflated catch statistics reported environment. At the Fisheries Ministry the coast- by China, the world’s biggest fishing nation. This guard and fisheries officers plan and execute moni- was thought to be due to local managers being toring programmes and quota allocations. On the under pressure to show increased production to other side of Vågen can be found the Bergen Fish- meet the goals of a centralized communist eries Museum as well as the warehouses and facto- economy. FAO estimates that 47–50% of the ries of fish processing and distribution firms. On world’s fish stocks are fully exploited, 15–18% the southern edge of the city centre, scientists in overexploited and 9–10% depleted (FAO 2000). A the Department of Fisheries and Marine Biology drop in fisheries production in 1998 was primarily of the University of Bergen are researching the due to the El Niño event that took place in 1997–8. ecology of marine organisms, work that underpins This influenced most directly the Southeast the conservation of biodiversity. This area of Pacific region, one of the regions that contributes 2 Chapter 1

140 Aquaculture 120 Capture fisheries

100

80

60 Million tonnes Fig. 1.1 Global landings of marine 40 and inland capture fisheries and aquaculture. Data from 1970 20 corrected for misreporting of marine landings (Watson and Pauly, 2001; R. Watson, personal 0 communication). (Source: from 1950195319561959196219651968 197119741977 19801983 1986198919921995 1998 FAO 2000.)

2.3 most to the catch of marine fish, mainly anchovy Altantic, Southwest 2.5 (Engraulis ringens) and Chilean jack mackerel 1998 (Trachurus murphyi) (Fig. 1.2). 2.8 1996 Pacific, Northwest As shown in Table 1.1, the catch of fish taken by 2.9 capture fisheries in inland waters continued to Atlantic, 3.6 grow between 1994 and 1999. In 1999 about 27 mil- Eastern Central 3.5 lion fishers landed 92 million tonnes of fish and shellfish. A further 9 million people were busy pro- Indian Ocean, 3.9 ducing 32 million tonnes of farmed fish. Interna- Western 4.1 tional trade in fisheries commodities was worth Indian Ocean, 4.0 some US$53.4 billion in 1999. The marine fishing Eastern 3.9 fleet consisted of around 23014 vessels over 100 5.0 tonnes. In reality there are probably at least as Asia, Inland waters many vessels again under this quite large lower 4.5 weight limit. The fish caught by all these vessels 8.0 Pacific, Southwest and people were processed and sold in a variety of 17.1 ways. In 1998, the latest year for which data are 9.3 available, 79.6% of fish landed was used for direct Pacific, Western Central 8.8 human consumption while the remainder went into non-food production. Of the proportion being 10.9 Atlantic, Northwest used directly as human food, 45.3% was sold fresh, 11 28.8% was frozen, 13.9% was canned and 12% 24.8 was cured. Fifty years ago the proportion that was Pacific, Northwest 23.5 frozen would have been much smaller and the part cured or canned much higher. 0 51015202530 Fish form an important part in the diet of many Million tonnes people. Average consumption per head has grown considerably over the past 40 years, increasing Fig. 1.2 Capture fisheries production by major fishing from 9kg per person per year in the early 1960s to regions in 1996 and 1998. (Source: from FAO 2000.) Underpinning Fisheries with Basic Biology 3

Table 1.1 Recent trends in world fisheries production and utilization. The data in this table are given without correction for the inflated catches documented by Watson and Pauly (2001). (Source: from FAO 2000.)

1994 1995 1996 1997 1998 1999

Production (million tonnes) Inland Capture 6.7 7.2 7.4 7.5 8.0 8.2 Aquaculture 12.1 14.1 16.0 17.6 18.7 19.8 Total inland 18.8 21.4 23.4 25.1 26.7 28.0 Marine Capture 84.7 84.3 86.0 86.1 78.3 84.1 Aquaculture 8.7 10.5 10.9 11.2 12.1 13.1 Total marine 93.4 94.8 96.9 97.3 90.4 97.2 Total capture 91.4 91.6 93.5 93.6 86.3 92.3 Total aquaculture 20.8 24.6 26.8 28.8 30.9 32.9 Total world fisheries 112.3 116.1 120.3 122.4 117.2 125.2

Utilization Human consumption (million tonnes) 79.8 86.5 90.7 93.9 93.3 92.6 Reduction to fishmeal and oil (million tonnes) 32.5 29.6 29.6 28.5 23.9 30.4 Human population (¥109) 5.6 5.7 5.7 5.8 5.9 6.0 Per-capita food fish supply (kg) 14.3 15.3 15.8 16.1 15.8 15.4

16kg per person per year in 1997 (FAO 2000; Table be taken into account at all times. Uncertainty 1.1). In developed countries consumption per head comes in different forms and has been categorized over the same period has risen from 19.7 to 27.7kg by Charles (1998) into three types: (i) random fluc- per year. Countries at the other end of the wealth tuations, (ii) uncertainty in parameters and states scale have much less fish available, although they of nature and (iii) structural uncertainty. The latter too have experienced a rise in supply from 4.9 to can be put more bluntly as ignorance; we just do 7.8kg per person per year. In poor countries fish not know how the system works. Ecology is a dif- can form 20% of a person’s protein intake. African ficult subject, especially when the species under countries have access to much less fish than do study is not easy to see and its population dynam- countries in Southeast Asia (excluding China). ics display unpredictable fluctuations, a phenome- non widespread in the natural world. Uncertainty in parameters and in the state of nature can now be 1.3 THE QUEST embraced by various estimation procedures, rang- FOR KNOWLEDGE ing from simulation methods to Bayesian statis- tics to fuzzy logic (Wade 2001). When all else fails, When T.H. Huxley made his infamous remarks in we may have to tackle structural uncertainty by 1883 about fish being too fecund for their numbers learning more about the system. ever to be influenced by fishing, he was showing The first volume of this book reviews the extent supreme confidence in his knowledge of fish and of our knowledge about many aspects of fish their biology. This confidence meant that others biology, while the second volume integrates this had to work hard to collect information before he knowledge with descriptions of a wide range of could be proven wrong (see Smith, Chapter 4, Vol- topics in fisheries biology and management, from ume 2). We now realize that uncertainty about fish how fish are caught to methods for assessing their biology and fishing systems is inherent and has to populations and predicting impacts of exploitation 4 Chapter 1 on the targeted species and their ecosystems. In ef- These two volumes show that we do know a great fect, the first volume tries to show what we know deal about fish biology and ecology, yet the world’s and to demarcate the limits of structural uncer- fish stocks are still being reduced mercilessly. The tainty about basic biology. Along the way the au- problem is that knowing about the resource is thors also show the degree to which our knowledge not enough. As Walters (1986) has said, fisheries is plagued by the other two sources of uncertainty. management is more about managing people than Some of the chapters inevitably overlap and the ecosystems, a point that we return to in our intro- different viewpoints on the same part of nature are duction to the second volume. reminiscent of a passage from John Steinbeck’s With two volumes composed of 34 chapters (1952) foreword to Between Pacific Tides by his written by 54 people, we hope we have provided a friend Ed Ricketts. He wrote: ‘There are good holistic view of fisheries science. In this chapter things to see in the tidepools and there are exciting we want to draw attention to our reasons for in- and interesting thoughts to be generated from the cluding the chapters we have in the first volume. In seeing. Every new eye applied to the peep hole addition we also try to show how the study of com- which looks at the world may fish in some new mercial fish populations has led to new insights beauty and some new pattern, and the world of the in many basic areas of biology and ecology. The human mind must be enriched by such fishing.’ early years of fisheries science created a discipline We think that these different viewpoints are that was considered very separate from the rest important as they illustrate how there is often of biology (Graham 1956). The need to tackle not enough known of an ecosystem or species to practical problems often made it hard for fisheries allow a definitive account to be given. scientists to contribute to the literature on basic Structural uncertainty is meat and drink to ecology. Despite this, scientists such as Hjort, the scientist. Problems are rarely solved entirely, Hardy, Beverton and Holt (see Smith, Chapter 4, because we are always uncovering new layers of Volume 2) left a lasting legacy in ecology as well as complexity. The view is often expressed that we in the specialist field of fisheries. really need to do more research on this or that problem so that we can better understand what is going on and therefore be in a better position to 1.4 PART 1: BIODIVERSITY manage it. To a point this is true, but as Ludwig et al. (1993) wrote: This volume starts with two accounts that tell us what fish there are, how we determine which Recently some of the world’s leading ecolo- taxa are related to one another, and where they gists have claimed that the key to a sustain- can be found (Gill and Mooi, Chapter 2; Mooi and able biosphere is research on a long list of Gill, Chapter 3). These are fascinating studies in standard research topics in ecology. Such a their own right, and the information also under- claim that basic research will (in an unspeci- pins several aspects of fisheries science. For exam- fied way) lead to sustainable use of resources ple, recent work that attempts to predict which in the face of a growing human population species are likely to be vulnerable to exploitation may lead to a false complacency: instead has used phylogenetic information to guide com- of addressing the problems of population parisons of responses to fishing. Common respon- growth and excessive use of resources, we ses could be due to close taxonomic relatedness may avoid such difficult issues by spending and this component needs to be accounted for money on basic ecological research. explicitly (see Reynolds et al., Chapter 15, Volume 2). Without this systematic knowledge, it would There is a danger that those interested in fisheries not be possible to make the comparisons necessary management and conservation will not know to derive rules of thumb telling us which species when to stop in the quest for more knowledge. are likely to be vulnerable to exploitation and Underpinning Fisheries with Basic Biology 5 which will be robust. Similarly, Hutchings (Chap- the fish from exploitation during spawning may ter 7, this volume) makes comparisons of life- not be enough. history variation in well-defined taxonomic groups. Furthermore, Rickman et al. (2000) showed that variation in recruitment depends in part on fecundity, but this relationship only be- 1.5 PART 2: comes apparent when comparisons are made be- PRODUCTION AND tween populations of the same species. Good POPULATION STRUCTURE systematics is also necessary for the analysis of substocks within species as described by Ward Sustainable fishing is maintained by the growth (Chapter 9, this volume). Finally, data collection, and reproductive output of individual fish. Growth as described by Evans and Grainger (Chapter 5, Vol- is governed by physiological ability to cope with ume 2), and stock assessment (Sparre and Hart, environmental challenges. Some of the physiolog- Chapter 13, Volume 2) are only possible if species ical mechanisms that fish have for living in water can be identified. Many tropical fisheries suffer are described by Brix (Chapter 4, this volume). Fish from poor knowledge of the systematics of the ex- stocks are often divided into local subunits, with ploited fish. For example, the trawl fishery on Lake their own adaptations to local conditions. A good Malawi catches 177 different taxa, most of which example of this is provided by the different haemo- are unnamed (Turner 1995). We can say without a globin genotypes possessed by cod that live close doubt that all the methods described in Volume 2 to the Norwegian coast. Brix describes how these require good systematics at the species or stock genotypes are linked to temperature-sensitive level. oxygen-binding properties of the haemoglobin Knowledge of biogeography not only allows us molecule, which adapt fish to the conditions to know where important species are found but in which they live. The haemoglobin genotypes also shows up hotspots of diversity where conser- also have an influence on competitive behaviour vation measures may be particularly needed. (Salvanes and Hart 2000). The temperature sensiti- Biogeographical knowledge also defines the type vity of the haemoglobin genotypes links back to of fish fauna expected from different habitats. the comments made earlier about cod being at Polunin and Pinnegar (Chapter 14, this volume), their southern limit in the North Sea. Fish are fine- Persson (Chapter 15, this volume) and Jones et al. tuned to the particular conditions they face and (Chapter 16, this volume) deal with the dynamics knowledge of physiology will make it possible to of communities in different biogeographical predict which fish will suffer or benefit from a zones. warming of the seas. Similarly, the physiological Knowledge of contemporary distributions will and behavioural attributes of fishes will determine be increasingly useful as fisheries biologists are how they can respond to pollution in estuaries and required to pay more attention to the influence the nearshore environment (Jones and Reynolds of climate change. This process will influence 1997). the distribution of fish, and knowledge of present In the next chapter, Jobling (Chapter 5, this biogeographical limits makes it possible to iden- volume) discusses the processing of energy and tify sensitive species. An example is the North its consequences for growth. Jobling also describes Sea stock of cod (Gadus morhua), which is at the way in which a fish develops and the models the southern limit of the species’ distribution that can be used to predict growth. The model (O’Brien et al. 2000). Recruitment is poor in of growth that is used most often is the von warm years. Should the North Sea continue to get Bertalanffy equation, which pops up in many warmer it will become harder for the cod to main- analyses of stock production. One interesting de- tain its current distribution and the limited area velopment is by Schnute and Richards (Chapter 6, closures employed in the spring of 2001 to protect Volume 2), who show how a generalized version of 6 Chapter 1 this model can also be used to describe population process or evolutionary responses mediated by growth. gene frequency changes. From individual productivity we move on to One aspect of life histories reviewed by population productivity. Myers (Chapter 6, this Hutchings (Chapter 7, this volume) concerns volume) describes the pattern of recruitment in a relationships between characteristics such as body range of species and discusses what is known about size, body growth rate, natural mortality rate and the causes of variation. This is a key question in length of first reproduction. These relationships fisheries science (Cushing 1996). Myers’ chapter were first studied by Beverton and Holt (1959) and underpins Chapters 6–13 in Volume 2, which deal have since been formalized and extended to other with aspects of stock assessment and modelling. taxa by Charnov (1993). These life-history invari- A key issue for the conservation of stocks and for ants have proved useful for predicting the vulnera- rebuilding those that have been overexploited is bility of populations to fishing (Pope et al. 2000). whether stocks at low abundance have the ability Likewise, Roff (1981, 1984) did pioneering work to produce enough offspring to regenerate the on the lifetime allocation of energy to growth and stock (compensatory reserve). An analysis of 246 reproduction and concluded in the earlier paper fish populations has shown that the maximum for the Pleuronectidae that reproductive lifespan annual reproductive rate, determined at the origin within the group was not a response to variations of the stock–recruit curve, is between 1 and 7. This in reproductive success but to variations in the age assumes that the stock–recruit curve goes through at maturity. Researchers such as Charnov (1993) the origin and implies that there are no disruptions and Charlesworth (1994) have since developed to the reproductive process as the spawning stock these early studies of aspects of life histories into becomes very small (the Allee effect). Frank and modern life-history theory, which is applicable Brickman (2000) showed that this Allee effect to a wide range of sexually reproducing species. can exist at the subpopulation level and this may Life histories and recruitment feed into migra- be obscured if data are aggregated from several tion, reviewed by Metcalfe et al. (Chapter 8, this subgroups. volume). Indeed, this chapter shows that migra- The study of life histories (Hutchings, Chapter tion is a fundamental facet of the life histories 7, this volume) follows naturally from studies of of many commercially important fish species. In recruitment. For example, an important link be- diadromous species, the life history has become tween fishing and recruitment is the size and age at split between two habitats, one of which is better which fish become sexually mature. These depend for feeding and the other for reproduction. Salmon on growth rates and other components of life histo- species provide classic cases of this lifestyle. Mi- ries that schedule reproduction in relation to the gration also makes them particularly vulnerable to lifespan. A major determinant of this schedule is fishing, because they must pass through a narrow the stage at which the animal experiences the most bottleneck on their way to spawn, and to environ- severe mortality. Law (1979) showed theoretically mental degradation through silting of spawning how increasing the mortality at some particular grounds and general deterioration in water quality age would lead to the evolution of increased repro- (see Reynolds et al., Chapter 15, Volume 2). ductive output in the years before that mortality Information on migration is essential for the occurred. This may lead to a reduced age at management of many stocks. The movements of maturity. A recent review by Reynolds et al. (2001) the fish will determine where the fishing boats has shown that large late-maturing fish are the will be. The way in which fishing effort is applied most vulnerable to exploitation. Thus, it is not sur- to the stock will be determined by this spatial cor- prising that other studies discussed by Hutchings relation and should be incorporated into models have shown that the principal life-history changes (see Sparre and Hart, Chapter 13, Volume 2). An due to fishing are reduced size and age at maturity. understanding of animal movements is also funda- These may be plastic responses of the growth mental to the success or failure of marine pro- Underpinning Fisheries with Basic Biology 7 tected areas (see Polunin, Chapter 14, Volume 2). implies considerable behavioural complexity. De- For example, the closure of 40% of the North Sea to tails such as this can be important in understand- cod fishing in the spring of 2001 is a sign that very ing spatial vulnerability of individuals, as well large areas have to be enclosed if a marine pro- as reproductive success. For example, Chapter 10 tected area is designed to conserve a species (this volume) shows how fertilization mode, aggre- that migrates over large distances. Indeed, many gation behaviour, sex changes and sperm limita- biologists are sceptical about whether even this tion all determine vulnerability to Allee effects area of closure was sufficient to have had appreci- (depensation). In the many reef species that under- able conservation impacts on the stock. go sex change, impacts of fishing on the popula- From migration we move on to a review of tions depend on the cues that determine sex population genetic structure of fish stocks (Ward, change. Fisheries typically remove the largest Chapter 9, this volume). Stocks used to be identi- individuals, and in many wrasses and groupers fied by morphological characters such as the num- these are males that had been females when they ber of vertebrae or the number of fin rays in dorsal were younger. If sex change is flexible according to or other fins, in conjunction with tagging studies. the females’ perceptions of presence or absence of Ward reviews the modern molecular methods that males, the loss of males to the fishery may be re- are now used, often in conjunction with advanced placeable. But if sex change occurs at a fixed size, technology to track fish migrations (see Metcalfe males may not be replaced fast enough, and sperm et al., Chapter 8, this volume). The increasing evi- depletion is possible. dence of the importance of substock structure in fish populations provides a growing market for this information. Molecular methods are also helping to determine relationships among taxa that can 1.6 PART 3: FISH AS then be compared to understand population char- PREDATORS AND PREY acteristics while taking into account phylogenetic relationships (Reynolds et al. 2001). Ward also Three chapters on feeding ecology illustrate the describes uses for new molecular techniques, such truth of Steinbeck’s observation about how differ- as identifying fishes or fish parts in markets, and ent people see the world in different ways depend- identifying endangered species offered for sale ing upon which ‘peep hole’ they use. Mittlebach’s in disguised form. A recent guide to Australian view in Chapter 11 (this volume) of predator–prey seafood contains a cellulose acetate-based finger- relationships is very much influenced by the theo- print for each of 380 species. retical structure that he employs to understand The final chapter in Part 2 rounds out the re- trophic ecology and habitat use. He is interested in views of various aspects of production and pop- explaining the ways that interactions between fish ulation structure with a review of reproduction species in freshwater lakes can be predicted from (Forsgren et al., Chapter 10, this volume). Tradi- a knowledge of the costs and benefits of different tionally, models in fisheries science have assumed prey types. The view of Juanes et al. in Chapter 12 that behavioural aspects of reproduction are not (this volume) focuses more clearly on the perspec- important. Cushing (1968) in his book on fisheries tive of fish as predators, with most attention paid biology devotes just one sentence to the spawning to the way in which feeding habits shape the life behaviour of the Norwegian cod in Vest Fjord, in histories, body form and ecology of predators. the Lofoten Islands: ‘The male fish arrive in the Krause et al. in Chapter 13 (this volume) take Vest Fjord first, and spawning takes place in the the prey’s point of view, emphasizing how their midwater layer where the fish are caught by drift ecology and behaviour are moulded by predation nets, long lines and purse seines . . .’. Subsequent pressure. work has led Nordeide and Folstad (2000) to All three foraging chapters take advantage of argue that cod may have a lekking system, which huge advances in our understanding of foraging 8 Chapter 1 theory through the application of optimization (Scomber scombrus) and herring (Clupea haren- methods (Stephens and Krebs 1986; Giraldeau and gus). Schooling behaviour also makes catch per Caraco 2000). They feed directly into the trophic unit effort (CPUE) useless for estimating the abun- models of fisheries, reviewed in the second volume dance of the stock as high CPUE is observed until by Pauly and Christensen (Chapter 10, Volume 2), the stock is virtually extinct. The catch rate of as well as multispecies virtual population analysis large aggregations of cod off Newfoundland re- reviewed by Shepherd and Pope (Chapter 7, mained constant at 1.5% of all cod catches right up Volume 2). For example, details of predatory until 1992 when the stock collapsed (Hutchings behaviour determine the shape of the ‘functional 1996). response’ of diet choice in relation to availability of alternative prey types (see Juanes et al., Chapter 12, this volume). The shape of this relationship is 1.7 PART 4: important in multispecies models of fisheries. FISH IN ECOSYSTEMS The basic research into prey choice and habitat choice reviewed by our triumvirate of chapters The final four chapters of this volume take the also feeds into our understanding of the role of details of behavioural interactions involving fish piscivores in trophic webs (Chapters 14–16, this as predators and prey (Part 3) and scale them up volume). For example, in freshwater systems, to understand community ecology. Polunin and predation on planktivorous fish species can lead to Pinnegar (Chapter 14) focus on marine food trophic cascades. This information has been used webs, Persson (Chapter 15) examines freshwater to improve water quality in many freshwater communities and Jones et al. (Chapter 16) make ecosystems through biomanipulation. Although it explicit comparisons between these habitats. An is unlikely that marine systems could be managed important theme that emerges is that as fish grow in the same way because of differences in scale, their niche changes so much that the young and fishers have changed the structure of marine food adults inhabit different worlds. Thus, in Swedish webs by reducing the abundance of large organ- lakes the interactions between European perch isms in much the same way as piscivores reduce (Perca fluviatilis) and roach (Rutilus rutilus) are a the abundance of planktivores (see Pauly and mixture of competition and predation that depend Christensen, Chapter 10 and Kaiser and Jennings, on the age and size of the interacting fish. A similar Chapter 16, Volume 2). Another market for the relationship holds for cod and gobies (Gobiusculus information on foraging behaviour (Chapters 11 flavescens) in Norwegian fjords. When cod are and 12, this volume) is in stocking programmes small they compete for food with gobies but gobies designed to boost freshwater sport fisheries (see are eaten by larger cod (A.G.V. Salvanes, personal Cowx, Chapter 17, Volume 2). We need to be able communication). The same is true of the large- to predict the likely benefits for anglers and the mouth bass (Micropterus salmoides) and blue- environmental damage that may be caused by gill sunfish (Lepomis macrochirus) in North stocking programmes (see Reynolds et al., Chap- American lakes. This information is critical for ter 15, Volume 2). The chapter by Krause et al. the implementation of multispecies models (see (Chapter 13, this volume) looks specifically at the Shepherd and Pope, Chapters 7 and 8, Volume 2) antipredator adaptations of fish that can lead to a and ecosystem models (see Pauly and Christensen, species being very vulnerable to fishing. For exam- Chapter 10, Volume 2). ple, schooling behaviour, which is so effective The final chapter of the first volume presents a against the attacks of individual fish predators, is detailed examination of fish parasites (Barber and a disadvantage in the face of fishing. Schooling Poulin, Chapter 17). The mixture of pure and ap- makes it relatively easy for purse seiners (Misund plied research reviewed here includes evidence et al., Chapter 2, Volume 2) to find fish using long- that parasites can change behaviour and drive range sonar and to catch species such as mackerel population events. For example, threespine stick- Underpinning Fisheries with Basic Biology 9 lebacks infected with Schistocephalus solidus are of Volume 1 of the Handbook is to provide the bio- more likely to take risks when feeding near preda- logical basis for the tools of management and con- tors than are uninfected individuals (Milinski servation described in Volume 2. Along the way, 1984). Likewise Milinski (1982) found that com- we have allowed ourselves (and our authors) the petitive interactions between sticklebacks were luxury of following paths that have led through altered by parasite load. That parasites seem able a considerable amount of additional background to drive population cycles was shown by Kennedy information that may never find its way directly et al. (1994) for Ligula intestinalis infecting roach into fisheries management, but which enriches in a lake in Devon, UK. For marine populations, our overall understanding of fish biology. We will there is little evidence for this type of popula- never reach the limits of our ignorance about the tion effect, although Holst et al. (1997) have shown way the fish world works but in the context of that there is a possible link between abundance management we do not need to know everything. of Norwegian spring-spawning herring (Clupea How much do we need to know? Ludwig et al. harengus) and infections of the fungal parasite (1993) present the case that it is silly to always Ichthyophorus sp. argue that we cannot act to manage stocks sustain- Parasites in fish can endanger human health. ably until we have more information. However, Herring often contain the nematode worm we cannot divorce the management of fishers from Anisakis simplex. The preparation of pickled her- the biology of their prey. As fish stocks have ring, particularly popular in Scandinavia and the collapsed, the search for solutions has led to new Netherlands, does not always kill the parasite in questions. As Hutchings (2000) has noted, many the fish flesh and the worm can be passed on to the of these concern ecology and evolution, subjects human consumer, sometimes finding its way to that are reviewed in this first volume. Hutchings the brain. Such infections can seriously under- draws examples of questions from the collapse and mine marketing efforts to persuade people to eat slow recovery of Atlantic cod along the eastern more fish (see Young and Muir, Chapter 3, Volume Canadian coast, to which we can add a few of 2). our own. Could the slow rate of recovery be due One of the biggest problems on salmon farms to changes in mating behaviour and Allee effects in Norway and Scotland has been the sea louse (depensation) at small population sizes? What (Lepeophtheirus salmonis). Estimated costs to the new predator–prey relationships might be in place industry vary widely but are likely to exceed £20 now that the fish occur at less than 5% of their million a year in both Scotland and Norway. Low population biomass of 30 years ago? Have eco- fish growth, stock losses and a reduced price at har- system shifts occurred that hinder recruitment? vest cause these costs. In addition the farmer has to What is the role of habitat disturbance? Are pay for monitoring and treatment, which can have temperature changes important? Have there been severe environmental impacts. At present, efforts evolutionary impacts on the populations? How are being made to use wrasse (Labridae) to act as many populations are there, and what kind of cleaners so reducing the need to use chemical source–sink dynamics might be involved in popu- treatments. The study of parasites is one area in lation recovery? fisheries science where there is little argument Few of these questions were being asked in any over the need for basic biological information. fishery 10 or 15 years ago. They are common now, and we do need the answers for effective manage- ment of many fish populations and ecosystems. How large should a marine reserve be and how 1.8 IGNORANCE close should it be to other reserves (see Polunin, BANISHED? Chapter 14 and Reynolds et al., Chapter 15, Vol- ume 2)? How will our management of one species As this chapter has outlined, the principal purpose affect populations and yields of others (see Pauly 10 Chapter 1 and Christensen, Chapter 10 and Kaiser and analytical methods, management priorities and the Jennings, Chapter 16, Volume 2)? How can we esti- Canadian groundfish experience. Fisheries Research mate natural mortality rates of target species, in 37, 37–50. Charlesworth, B. (1994) Evolution of Age-structured order to make accurate forecasts from models (see Populations. Cambridge: Cambridge University Press. Shepherd and Pope, Chapter 7, Volume 2)? How do Charnov, E.L. (1993) Life History Invariants: Some life histories, behaviour and economic value affect Explorations of Symmetry and Evolutionary Ecology. vulnerability to extinction (see Reynolds et al., Oxford: Oxford University Press. Chapter 15, Volume 2)? Cushing, D.H. (1968) Fisheries Biology. A Study in Popu- Many of the contributors to these two volumes lation Dynamics. Madison: University of Wisconsin have never worked alongside one another before. Press. Cushing, D.H. (1996) Towards a Science of Recruit- We hope that these two volumes will help to bridge ment in Fish Populations. Oldendorf/Luhe: Ecology the gap between pure and applied research, leading Institute. to a productive dialogue between those who study FAO Fisheries Department (2000) The state of world fish biology for different reasons but whose inter- fisheries and aquaculture. Rome: Food and Agricul- ests may have more in common than they realize. tural Organization of the United Nations. http:// www.fao.org/sof/sofia/index_en.htm Frank, K.T. and Brickman, D. (2000) Allee effects and compensatory population dynamics within a stock 1.9 CONCLUSIONS complex. Canadian Journal of Fisheries and Aquatic Sciences 57, 513–17. We have tried in this chapter to illustrate how Giraldeau, L.-A. and Caraco, T. (2000) Social Foraging the material in the 17 chapters of Volume 1 of this Theory. Princeton, NJ: Princeton University Press. Graham, M. (ed.) (1956) Sea Fisheries: Their Inves- Handbook can be used for an intelligent applica- tigation in the United Kingdom. London: Edward tion of the methods covered in Volume 2. Fisheries Arnold. scientists are often accused of being so preoccu- Holst, J.C., Salvanes, A.G.V. and Johansen, T. (1997) pied with the details of stock assessment that they Feeding, Ichthyophorus sp. infection, distribution and forget about individual interactions by fish within growth history of Norwegian spring-spawning herring stocks as well as connectivity between stocks and in summer. Journal of Fish Biology 50, 652–64. the rest of the ecosystem. We hope that the infor- Hutchings, J.A. (1996) Spatial and temporal variation in the density of northern cod and a review of hypotheses mation in this volume will be of interest for its for the stock’s collapse. Canadian Journal of Fisheries own sake, and for helping to put ecology and evolu- and Aquatic Sciences 53, 943–62. tion back into fishery science and management. Hutchings, J.A. (2000) Numerical assessment in the front seat, ecology and evolution in the back seat: time to change drivers in fisheries and aquatic sciences? Marine Ecology Progress Series 208, 299–303. ACKNOWLEDGEMENTS Jennings, S., Reynolds, J.D. and Mills, S.C. (1998) Life history correlates of responses to fisheries exploita- We thank Anne Gro Salvanes for valuable com- tion. Proceedings of the Royal Society of London Series ments on an earlier draft of this chapter. B 265, 333–9. Jones, J.C. and Reynolds, J.D. (1997) Effects of pollution on reproductive behaviour of fishes. Reviews in Fish Biology and Fisheries 7, 463–91. REFERENCES Kennedy, C.R., Wyatt, R.J. and Starr, K. (1994) The decline and natural recovery of an unmanaged coarse Beverton, R.J.H. and Holt, S.J. (1959) A review of the fishery in relation to changes in land use and attendant lifespans and mortality rates of fish in nature, and their eutrophication. In: I.G. Cowx (ed.) Rehabilitation of relation to growth and other physiological characteris- Freshwater Fisheries. Oxford: Fishing News Books, tics. CIBA Foundation Colloquia on Ageing 54, 142– pp. 366–75. 80. Law, R. (1979) Optimal life histories under age-specific Charles, A.T. (1998) Living with uncertainty in fisheries: predation. American Naturalist 114, 399–417. Underpinning Fisheries with Basic Biology 11

Ludwig, D., Hilborn, R. and Walters, C. (1993) Uncer- Roff, D.A. (1981) Reproductive uncertainty and the tainty, resource exploitation, and conservation: evolution of interoparity: why don’t flatfish put all lessons from history. Science 260, 17–36. their eggs in one basket? Canadian Journal of Fisheries Milinski, M. (1982) Optimal foraging: the influence of and Aquatic Sciences 38, 968–77. intraspecific competition on diet selection. Behav- Roff, D.A. (1984) The evolution of life-history para- ioral Ecology and Sociobiology 11, 109–15. meters in teleosts. Canadian Journal of Fisheries and Milinski, M. (1984) Parasites determine a predator’s Aquatic Sciences 41, 989–1000. optimal feeding strategy. Behavioral Ecology and Salvanes, A.G.V. and Hart, P.J.B. (2000) Is individual Sociobiology 15, 35–7. variation in competitive performance of reared juve- Nordeide, J.T. and Folstad, I. (2001) Is cod lekking or a nile cod influenced by haemoglobin genotype? Sarsia promiscuous group spawner? Fish and Fisheries 1, 85, 265–74. 90–3. Steinbeck, J. (1952) Foreword. In: E.F. Ricketts and O’Brien, C.M., Fox, C.J., Planque, B. and Casey, J. (2000) J. Calvin Between Pacific Tides, 3rd edn (revised by Fisheries: climate variability and North Sea cod. Na- J.W. Hedgpeth). Stanford, CA: Stanford University ture 404, 142. Press. Parrish, B.B. (1956) The cod, haddock and hake. In: M. Stephens, D. and Krebs, J.R. (1986) Foraging Theory. Graham (ed). Sea Fisheries: Their Investigation in Princeton, NJ: Princeton University Press. the United Kingdom. London: Edward Arnold, pp. Turner, G.F. (1995) Changes in demersal cichlid commu- 251–331. nities as a result of trawling in southern Lake Malawi. Pope, J.G., MacDonald, D.S., Daan, N., Reynolds, J.D. In: T.J. Pitcher and P.J.B. Hart (eds) The Impact of and Jennings, S. (2000) Gauging the impact of fishing Species Changes in African Lakes. London: Chapman mortality on non-target species. ICES Journal of & Hall, pp. 387–412. Marine Science 57, 689–96. Wade, P.R. (2001) The conservation of exploited species Reynolds, J.D., Jennings, S. and Dulvy, N.K. (2001) Life in an uncertain world: novel methods and the failure of histories of fishes and population responses to ex- traditional techniques. In: J.D. Reynolds, G.M. Mace, ploitation. In: J.D. Reynolds, G.M. Mace, K.H. Redford K.H. Redford and J.G. Robinson (eds) Conservation of and J.G. Robinson (eds) Conservation of Exploited Exploited Species. Cambridge: Cambridge University Species. Cambridge: Cambridge University Press, pp. Press, pp. 110–43. 147–68. Walters, C.J. (1986) Adaptive Management of Rickman, S.J., Dulvy, N.K., Jennings, S. and Reynolds, Renewable Resources. New York: Macmillan. (Also J.D. (2000) Recruitment variation related to fecundity reprinted by the Fisheries Centre, University of British in marine fishes. Canadian Journal of Fisheries and Columbia, Vancouver, BC, Canada.) Aquatic Sciences 57, 116–24. Watson, R. and Pauly, D. (2001) Systematic distortions in Righton, D., Metcalfe, J. and Connolly, P. (2001) world fisheries catch trends. Nature 414, 534–6. Fisheries: different behaviour of North and Irish sea cod. Nature 411, 156. Part 1 Biodiversity 2 Phylogeny and Systematics of Fishes

A.C. GILL AND R.D. MOOI

2.1 INTRODUCTION 2.2 PHYLOGENETIC METHODS AND The term ‘fish’, in the conventional sense, refers CLASSIFICATION to craniate animals (i.e. hagfishes and vertebrates) that are aquatic, with permanent gills borne on Over the last 40 years, three main methods have pharyngeal pouches or arches, median fins sup- vied for support as techniques that can be used to ported by cartilaginous or bony rays, and usually classify and infer relationships among organisms: paired fins but never limbs bearing digits. Fishes (i) numerical taxonomy or phenetics; (ii) evolu- represent a diverse group that includes over 25000 tionary taxonomy; and (iii) cladistics or phyloge- extant species, or roughly half of the Craniata (J.S. netic systematics. Space constraints do not allow a Nelson 1994; Helfman et al. 1997; Lundberg et al. detailed discussion or comparison of these meth- 2000). However, as so defined, they do not consti- ods; we offer only a brief contrast, mainly to aid the tute a ‘natural’ or monophyletic group (see below) reader in interpreting our discussion of fish orders because they do not share a common ancestor ex- and their interrelationships. Numerous publica- clusive of the Tetrapoda (reptiles, birds, mammals tions discuss and contrast the three methods in and amphibians). much greater detail (e.g. Hennig 1966; G.J. Nelson Systematics, or phylogenetic systematics, is 1972; Nelson & Platnick 1981); much of the debate the field of biology involved with the investigation regarding the aims of systematics was played out of phylogenetic relationships and classification of in the journal Systematic Zoology (now Syste- organisms. Cladistic methods have come to domi- matic Biology) during the 1970s. Moreover, debate nate systematic investigation over the past few about methods in cladistics continue, particularly decades, and for this reason, as well as our own in the journal Cladistics (journal of the Willi subscription to this method, we write this chapter Hennig Society). The interested reader should from a cladistic perspective. consult these references. The aims of this chapter are (i) to outline meth- The three methods differ fundamentally in how ods of inferring phylogenetic relationships among characters are analysed and how taxa are classified organisms and of classification focusing on cladis- (Fig. 2.1). Characters are the observable parts or at- tics; and (ii) to provide an overview of higher fish tributes of organisms, and are effectively the units taxa that includes brief details of taxonomic diver- of systematic analysis. Obviously, in order that sity, habits and general distribution, and to direct characters may be used to differentiate taxa, they the reader to literature that details relationships must vary; that is, they must express two or more within and among such taxa. character states. Numerical taxonomy is based on a concept of overall similarity, whereupon rela- 16 Chapter 2 tionships are inferred without regard to whether This conclusion is based on a concept of parsi- character states are derived (apomorphic) or primi- mony: in the absence of evidence to the contrary, tive (plesiomorphic). For example, given the sim- such as a conflict with other characters (see below), ple matrix in Fig. 2.1, a pheneticist would infer a it is more parsimonious to conclude that a de- close relationship between taxa B and C, because rived character state has evolved once in a com- these taxa possess the same character states for mon ancestor with inheritance by descendant taxa two characters (Y and Z), whereas only one charac- than to conclude multiple independent origins of ter (X) suggests an alternative close relationship the apomorphy. The shared presence of primitive between taxa A and B. A resulting phenetic classi- (plesiomorphic) characters (symplesiomorphy) fication would place taxa B and C in a more general neither refutes nor supports common ancestry. taxon that excludes taxon A. Complex statistical Moreover, the presence of apomorphic characters methods are typically used in phenetic studies and in a single taxon (autapomorphy) does not con- the development of phenetics coincided with the tribute to an understanding of that taxon’s rela- development of computer technology; in particu- tionships to other taxa. Thus, in the matrix given lar, cluster analyses are often used, usually in asso- in Fig. 2.1, only one derived character (X) is shared ciation with large arrays of characters under the by two taxa (A and B), leading to a conclusion that assumption that this leads to greater objectivity. these are sister taxa. The apomorphic states for Nevertheless, relationships generated by phenetic characters Y and Z are found only in taxon A, methods are concerned with general similarity and do not contribute to an understanding of its rather than genealogy, and have mostly fallen out relationships to the other taxa. However, notions of favour with systematists. They do persist in cer- of synapomorphy, autapomorphy and symple- tain fields, however, particularly in biochemical siomorphy are dependent on perspective, on the systematics, illustrated by analyses based on ge- level of generality of the analysis at hand. For netic distance. example, if we consider that taxon A is a family Cladistic methods, which were first formulated consisting of multiple species, the apomorphic by Hennig (1950; later revised and published in states (1 in Fig. 2.1) for characters Y and Z, which English, Hennig 1966), differ from phenetic we have already indicated are autapomorphies of A methods in giving importance only to derived in an analysis of the relationships of A, B and C, are (apomorphic) character states. The presence of a also synapomorphies uniting the various species given apomorphic state in two or more taxa in the family A together; but they are also symple- (synapomorphy) is evidence of common ancestry. siomorphies, and so uninformative, in an analysis of relationships among the species in A. Phylogenies are inferred from hierarchically distributed synapomorphies. However, in reality, ABC C B A there is almost always conflict among charac- ters; apparent synapomorphies may suggest dif- ferent phylogenetic relationships. Phylogenies are constructed, usually with the aid of computer algorithms, such that conflict between assumed Taxa X Y Z A 1 1 1 synapomorphies are minimized under a parsi- B 1 0 0 mony argument; thus trees with minimal length, C 0 0 0 which are those requiring the fewest character- (a) (b) state changes and therefore assumptions about convergence or homoplasy, or apparent loss of apo- Fig. 2.1 Matrix of three characters (X, Y and Z) for three taxa and the relationships inferred using (a) phenetic and morphies, are favoured. As such, cladistic analyses (b) cladistic methods. 0, Plesiomorphic character states; can also be viewed as tests of character-state ho- 1, apomorphic character states. mology: assumed synapomorphies that fail to diag- Phylogeny and Systematics 17 nose monophyletic groups (see below) are non- graphical, temporal and ontogenetic distribution. homologous; conversely, those that do diagnose We are also limited in our understanding by the monophyletic groups are homologous (Patterson historical basis of the characters under investi- 1982, 1988; de Pinna 1991; G.J. Nelson 1994). gation. Most systematists have concluded that Cladistic classifications differ from phenetic cladistic methods presently offer the best solution ones in that only monophyletic taxa (clades) are for both inferring phylogeny and constructing clas- recognized: these can be viewed as groups that sifications. Our treatment of fish classification in include an ancestor and all of its descendants. In this chapter reflects this cladistic approach. contrast, pheneticists may also recognize para- phyletic taxa (grades) in their classifications: these 2.2.1 Types of characters in are groups that exclude one or more of an ances- fish systematics tor’s descendants such as a taxon consisting of B and C in Fig. 2.1. In short, cladistic methods effec- For fishes, osteological and external characters tively equate the terms ‘synapomorphy’, ‘homo- have dominated historically, at least in part be- logue’, ‘relationship’ and ‘taxon’. cause they are most easily studied, particularly in The third method, evolutionary taxonomy, is fossil specimens. It is therefore not surprising that more difficult to describe because it is not a unified most higher taxa are diagnosed mainly or entirely method. Usually, evolutionary taxonomists em- by such characters. The modern discovery and ploy cladistic methods for inferring phylogenetic survey of osteological characters has been greatly relationships, though often invoking assumptions enhanced by the relatively recent development about the evolution of character states, but depart of improved methods for preparing specimens from cladistics in often recognizing paraphyletic (Taylor 1967; Dingerkus and Uhler 1977; Taylor taxa. Such taxa are recognized to reflect perceived and Van Dyke 1985). gaps in evolution. Thus, an evolutionary taxono- Muscles, particularly those operating the jaws, mist may infer the same relationships as a cladist have received a reasonable level of investigation for the example given in Fig. 2.1, but may conclude for character analyses, and have also played a role the same classification as a pheneticist (i.e. two in determining topographical relationships of os- taxa: A, B + C) if the autapomorphies of A were per- teological features, an important step in refining ceived as sufficient to place it in a different taxon our definition of such characters. Other systems from B + C. Recently, in discussing the classifica- associated with soft anatomy, such as neural tion of tunas and billfishes, Carpenter et al. (1995) anatomy, heart and blood vessel anatomy, have defended the use of evolutionary taxonomic not been as comprehensively studied, though they methods, arguing that they led to greater stability are of considerable potential, both as characters in of taxonomic nomenclature. their own right and for refining our understanding In short, the primary objectives of systematics of characters for osteological and other studies. For are to (i) discover characters; (ii) investigate their example, Parenti and Song (1996) studied innerva- distributions; (iii) determine the phylogenetic re- tion patterns in an attempt to define more care- lationships (hierarchical pattern) implied by (i) and fully the character ‘pelvic-fin position’, which (ii); and (iv) establish classifications that reflect the has had a long history in systematic ichthyology. relationships in (iii). These objectives represent Other morphological systems that have been significant challenges. Our understanding of char- important sources of characters include egg mor- acter distributions is often hampered by tradi- phology (e.g. White et al. 1984; Mooi 1990; Britz tional classifications and views of relationships as 1997), scale ultrastructure (e.g. Roberts 1993) illustrated in our discussion of acanthomorph and spermatozoan morphology (e.g. Jamieson fishes. In some sense assumptions are always 1991). Behavioural characters has also been em- made about the distribution of characters, because ployed in phylogenetic studies. For example, the we lack the resources to investigate their real geo- Ostariophysi are distinctive in possessing an 18 Chapter 2 alarm reaction to alarm substance (see below) Table 2.1 Classification of the Eurasian perch and McLennan et al. (1988) employed behavioural (Perca fluviatilis). characters to investigate relationships among Phylum Chordata gasterosteids. Class Actinopterygii Finally, recent years have seen a strong bias to- Order Perciformes wards the use of biochemical characters, such as Family Percidae isozyme and sequence data. Despite tremendous Tribe Percini technological advances in methods of, for exam- Genus Perca ple, DNA extraction and sequencing and their Species fluviatilis demonstrated utility at lower taxonomic levels, biochemical characters have yet to contribute in a significant way to our understanding of the higher relationships of fishes. This, in part, reflects our and species. Additional divisions are also often rec- poor understanding of biochemical characters, ognized within these by adding a ‘super’, ‘sub’ or their distribution and homology, in short the lack ‘infra’ prefix to these ranks. Examples are ‘super- of an appropriately comprehensive historical family’ between the order and family level, and basis. Studies of higher relationship in which one ‘subfamily’ between the family and tribe level. An or a few exemplars from huge clades are surveyed, alternative is to slot in less commonly used ranks, which is the current norm in molecular studies of such as ‘division’ and ‘cohort’. The hierarchical higher relationships of fishes, are unlikely to pro- structure of these ranks is convenient for reflect- duce a realistic understanding of characters, their ing nested phylogenetic relationships, but their distribution, or the relationships they imply. For use has resulted in considerable confusion about example, Rasmussen and Arnason’s (1999) conclu- the reality of rank. sion from an analysis of mitochondrial DNA se- The artificial nature of ranking is immediately quences that cartilaginous fishes are nested within obvious when one examines the phylogeny of bony fishes is flawed, because the eight fish species extant fish order provided in Fig. 2.2. Clearly the surveyed do not meaningfully represent 25000 or orders of fishes are not comparable, either in the so fish species. Similar concerns about taxonomic sense of relative age as indicated by their relative sampling were expressed by Hennig (1966, p. 103) position in the branching sequence, or in the sense in his discussion of early studies of serum char- of taxonomic diversity as perhaps indicated by the acters. Nevertheless, it is likely that this situ- number of families or species. They are a neces- ation will improve as our historical basis for sary, though uncomfortable, compromise in com- biochemical characters improves and more taxa municating concepts of phylogenetic relationship are surveyed. As these challenges are overcome, within the framework of traditional Linnaean clas- molecular characters will provide fish systema- sification. However, the observation that ranking tists with valuable information that, along with is arbitrary extends to all taxonomic levels, includ- continued analysis of morphological data, will ing – perhaps disturbingly for many biologists – significantly advance higher-level phylogeny species. The latter observation is at odds with construction. various practices involved with the tallying of species, usually in the manufacture of indices of 2.2.2 Taxonomic ranks and species diversity; because species are not comparable, such tallies allow, at best, only gross comparison A hierarchical system of taxa is currently used in although we continue this tradition below in dis- biological classification (Table 2.1). Thus, a phy- cussing such things as ‘sizes’ of orders. lum consists of one or more classes, each of which The species issue deserves further discussion, if consists of one or more orders, and so on through only in respect for the substantial baggage associ- less inclusive taxa such as families, tribes, genera ated with that rank. More so than any other taxo- Phylogeny and Systematics 19

iformes g thiformes lossiformes iformes teriformes cephalans

guill peiformes

MyxiniformesPetromyzontiformesChimaeriformesHeterodontiformesOrectolobiformesCarcharhiniformesLamniformesHexanchiformesEchinorhiniformesSqualiformesSquatiniformesPristiophoriformesRajiformesCoelacanthCeratodontiformesLepidosireniformesTetrapodaPolyp AcipenseriformesLepisosteiformesAmiiformesHiodontiformesOsteog ElopiformesAlbuliformesNotacanAn Clu GonorhynchiformesCypriniformesCharaciformesSiluriformesGymnotiformesRemainin clupeo

18 16 17

14 15 13 12

10 6 7 Fig. 2.2 Interrelationships 5 9 11 among extant fish orders. 4 8 3 Numbers refer to supraordinal 2 taxa: 1, Craniata; 2, Vertebrata; 1 3, Gnathostomata; 4, Chondrichthyes; 5, Elasmobranchii; 6, Galeomorphii;

beryciformes ontiformes 7, Squalea; 8, Osteichthyes; oidiformes ontiformes d iformes 9, Sarcopterygii; 10, Dipnoi; phiiformes o 11, Actinopterygii; 12, Teleostei; ArgentiniformesSalmoniformesEsociformesStomiiformesAteleopodiformesAulopiformesMyctophiformesLampridiformesPolymixiiformesPercopsiformesOphidiiformesGadiformesBatrachL StephanoZeiiformesBeryciiformesMugil AtheriniformesBeloniformesCyprinodSynbranchiformesGasterosteiformesDactylopteriformesGobiesociformesPerciformesPleuronectiformesTetrao 13, Osteoglossomorpha; 14, Elopomorpha; 27 15, Clupeocephala; 26 16, Clupeomorpha; 17, Ostariophysi; 18, Otophysi; 19, Protacanthopterygii; 25 20, Neoteleostei; 21, Eurypterygii; 22, Cyclosquamata; 23, Ctenosquamata; 22 24 24, Acanthomorpha; 23 21 25, Paracanthopterygii; 20 26, Smegmamorpha; 19 27, Atherinomorpha.

nomic rank, the species rank has been debated and about species. However, there is little consensus usually seen as having a special reality or proper- among biologists concerning species concepts, not ties that uniquely distinguish it from all other even among systematic biologists, the people who taxonomic ranks. For example, species are often routinely describe and revise species. For example, viewed as the ‘building blocks’ of biological diver- Mayden (1997) listed some 22 different species sity, the ‘fundamental units’ of evolution, unlike concepts (Table 2.2), which variously address dif- higher taxa, which are seen as the arbitrary units ferent interests and processes and thus attributes resulting from such evolution. Because of this of biological diversity. For example, the biological view that species have some sort of special reality, species concept places emphasis on the reproduc- many biologists feel qualified to make judgements tive interactions between individuals. 20 Chapter 2

Table 2.2 Species concepts compiled by Mayden (1997). can be tested and modified by the collection of more data and their congruence as interpreted by Agamospecies concept cladistic analysis. Biological species concept We do not expect the reader to agree with the Cladistic species concept Cohesion species concept view that species are merely taxa – the smallest Composite species concept taxa identifiable – and that taxa of the same rank Ecological species concept are not comparable. Regardless, we do hope that Evolutionary significant unit the reader will appreciate that species are not di- Evolutionary species concept rectly comparable given that divergent species Genealogical concordance concept concepts are in use. A recent summary of views on Genetic species concept Genotypic cluster definition the species concept in fish biology can be found in Hennigian species concept Nelson and Hart (1999). Internodal species concept Morphological species concept Non-dimensional species concept 2.3 FISH DIVERSITY Phenetic species concept Phylogenetic species concept AND PHYLOGENY Diagnosable version Monophyly version Phylogenetic relationships among fishes are per- Diagnosable/monophyly version haps better known than any other animal group of Polythetic species concept comparable size. This partly reflects the early and Recognition species concept important contributions of several ichthyologists Reproductive competition concept including G.J. Nelson, Patterson, Rosen and Wiley Successional species concept Taxonomic species concept in the development and popularizing of cladistic theory and methods, and of the general influence of these individuals on the ichthyological commu- nity. However, it also reflects the cumulative ef- fects of a very long history of endeavour in fish De Queiroz (1998) has argued that there is really classification and anatomy, and the important only one species concept, that of species as an evo- contributions of a long line of dedicated ichthyolo- lutionary lineage, and that the arguments con- gists. A small sample of these would include the cerning ‘concepts’ as listed by Mayden are really French anatomist Georges Cuvier (1769–1832), criteria, i.e. ways to identify whether or not a group the Swiss anatomist Louis Agassiz (1807–1873), of organisms is indeed a separate lineage. As sys- the American ichthyologists Theodore Gill tematists, we find G.J. Nelson’s (1989a; see also (1837–1914) and David Starr Jordan (1851–1931), Nelson and Hart 1999) argument compelling that and the British ichthyologists Albert Günther species are taxa and are nothing more nor nothing (1830–1914) and C. Tate Regan (1878–1943). less. Moreover, we believe that this concept, Equally important has been the associated devel- which places emphasis on pattern rather than opment of museum collections of both specimens process, will ultimately lead to discovery about and literature, which form the basis of our under- process. Although some systematists argue for the standing of fish systematics (Pietsch and Anderson recognition of subspecies and other intraspecific 1997). Despite this, however, there is much dis- ranks (e.g. Randall 1998), we for various reasons agreement among ichthyologists regarding even (some of which are summarized by Gill 1999; Gill interordinal relationships of fishes, particularly and Kemp, in press) argue against the use of in- among the more ‘advanced’ taxa, and the same traspecific ranks, and that species should be the holds true for the classification of fishes. least inclusive monophyletic taxa identifiable. As Comprehensive classifications of fishes are pro- such, species are hypotheses of relationships that vided in Greenwood et al. (1966, 1973), G.J. Nelson Phylogeny and Systematics 21

(1969), J.S. Nelson (1976, 1984, 1994), Lauder and other characters suggest that the lampreys are Liem (1983), Eschmeyer (1990), and Paxton and more closely related to the remaining extant crani- Eschmeyer (1994, 1998). Other publications that ates than they are to the hagfishes, with the main deal with the classification and relationships of evidence being the presence of embryonic or one or more major fish taxa include Moser et al. rudimentary vertebral elements, and cladistic (1984), Cohen (1989), Johnson and Anderson classifications place the lampreys in a mono- (1993), Stiassny et al. (1996) and Malabarba et al. phyletic Vertebrata that excludes the hagfishes (1998). (Janvier 1981). The following discussion of fish phylogeny and The Petromyzontiformes (lampreys) is a small classification (Fig. 2.2; Table 2.3) is not intended order with about 40 species of freshwater and to provide new hypotheses. Rather, the intention anadromous fishes found in cool temperate areas is to describe higher fish taxa based on recent around the world. One extant family (Petromyzon- literature, including aspects of taxonomic diver- tidae) is restricted to the Northern Hemisphere; sity, habits and general distribution, and to direct the remaining two extant families (Geotriidae the reader to literature that details interrelation- and Mordaciidae) are restricted to the Southern ships and intrarelationships and associated char- Hemisphere. There is disagreement about the rank acter evidence among such taxa. We have also of these taxa, with some authors recognizing a attempted to identify major areas of controversy single family (Petromyzontidae) with three extant and lack of resolution. subfamilies (e.g. J.S. Nelson 1994). The remaining extant fishes are classified in a 2.3.1 Extant fish orders and taxon called the Gnathostomata (jawed verte- brates). Character evidence for monophyly of this their interrelationships group includes presence of biting jaws derived The Myxiniformes (hagfishes) is a small order with from gill arches, spinal nerve roots united, semi- about 45 extant species of marine fishes found circular canal horizontal, and gill arches mainly in temperate waters throughout the world. jointed and medially placed (Janvier 1996). They are scavengers, feeding mainly on the insides Living gnathostomes are divisible into two main of dying or dead fishes and invertebrates. They are groups, Chondrichthyes (cartilaginous fishes) and sometimes classified with the lampreys (Petromy- Osteichthyes (bony fishes and tetrapods). zontiformes) in the superclass Agnatha, but the The Chondrichthyes is divided into two taxa: characters for this grouping are primitive and thus the Holocephali, with a single extant order not informative of a close relationship. Moreover, (Chimaeriformes), and the Elasmobranchii (sharks

Table 2.3 Classification of extant fishes. Orders and families are listed in phylogenetic and alphabetical sequences respectively. Commonly used synonyms of certain orders and families are indicated in parentheses. Families that include important commercial and recreational fisheries species are indicated in bold.

Myxiniformes [Myxinidae] Petromyzontiformes [Geotriidae, Mordaciidae, Petromyzontidae] Chimaeriformes [Callorhynchidae, Chimaeridae, Rhinochimaeridae] Heterodontiformes [Heterodontidae] Orectolobiformes [Brachaeluridae, Ginglymostomatidae, Hemiscylliidae, Orectolobidae, Parascylliidae, Rhincodontidae, Stegostomatidae] Carcharhiniformes [Carcharhinidae (Sphyrnidae), Hemigaleidae, Leptochariidae, Proscylliidae, Pseudotriakidae, Scyliorhinidae, Triakidae] Lamniformes [Alopiidae, Cetorhinidae, Lamnidae, Megachasmidae, Mitsukurinidae, Odontaspididae, Pseudocarchariidae] Hexanchiformes [Chlamydoselachidae, Hexanchidae] Echinorhiniformes [Echinorhinidae] Squaliformes [Centrophoridae, Dalatiidae, Squalidae] 22 Chapter 2

Table 2.3 Continued

Squatiniformes [Squatinidae] Pristiophoriformes [Pristiophoridae] Rajiformes [Dasyatidae (Gymnuridae, Myliobatidae, Mobulidae), Hexatrygonidae, Hypnidae, Narcinidae, Narkidae, Platyrhinidae, Plesiobatidae, Potamotrygonidae, Pristidae, Rajidae, Rhinidae, Rhynchobatidae (Rhinobatidae), Torpedinidae, Urolophidae, Urotrygonidae, Zanobatidae] Coelacanthiformes [Latimeriidae] Ceratodontiformes [Neoceratodontidae] Lepidosireniformes [Lepidosirenidae (Protopteridae)] Polypteriformes [Polypteridae] Acipenseriformes [Acipenseridae, Polyodontidae] Lepisosteiformes [Lepisosteidae] Amiiformes [Amiidae] Hiodontiformes [Hiodontidae] Osteoglossiformes [Gymnarchidae, Mormyridae, Notopteridae, Osteoglossidae (Pantodontidae)] Elopiformes [Elopidae, Megalopidae] Albuliformes [Albulidae, Pterothrissidae] Notacanthiformes [Halosauridae, Notacanthidae (Lipogenyidae)] Anguilliformes [Anguillidae, Chlopsidae (Xenocongridae), Colocongridae, Congridae, Cyematidae, Derichthyidae, Eurypharyngidae, Heterenchelyidae, Monognathidae, Moringuidae, Muraenesocidae, Muraenidae, Myrocongridae, Nemichthyidae, Nettastomatidae, Ophichthidae, Saccopharyngidae, Serrivomeridae, Synaphobranchidae] Clupeiformes [Chirocentridae, Clupeidae (Sundasalangidae), Denticipitidae, Engraulidae, Pristigasteridae] Gonorynchiformes [Chanidae, Gonorynchidae, Kneriidae, Phractolaemidae] Cypriniformes [Balitoridae (Gasteromyzontidae, Homalopteridae), Catostomidae, Cobitidae, Cyprinidae (Psilorhynchidae), Gyrinocheilidae] Characiformes [Acestrorhynchidae, Alestiidae, Anostomidae, Characidae, Chilodontidae, Citharinidae, Crenuchidae, Ctenoluciidae, Curimatidae, Cynodontidae, Distichodontidae (Ichthyboridae), Erythrinidae, Gasteropelecidae, Hemiodontidae (Anodontidae), Hepsetidae, Lebiasinidae, Parodontidae, Prochilodontidae] Siluriformes [Akysidae (Parakysidae), Amblycipitidae, Amphiliidae, Anchariidae, Andinichthyidae, Ariidae, Aspredinidae, Astroblepidae (Argidae), Auchenipteridae (Ageneiosidae, Centromochlidae), Austroglanididae, Bagridae (Olyridae), Callichthyidae, Cetopsidae (Helogeneidae), Chacidae, Clariidae, Claroteidae (Auchenoglanididae), Cranoglanididae, Diplomystidae, Doradidae, Erethistidae, Heteropneustidae (Saccobranchidae), Ictaluridae, Loricariidae, Malapteruridae, Mochokidae, Nematogenyidae, Pangasiidae, Pimelodidae (Heptapteridae, Hypophthalmidae, Pseudopimelodidae), Plotosidae, Schilbidae, Scoloplacidae, Siluridae, Sisoridae, Trichomycteridae] Gymnotiformes [Apteronotidae (Sternarchidae), Hypopomidae, Gymnotidae (Electrophoridae), Rhamphichthyidae, Sternopygidae] Argentiniformes [Alepocephalidae (Leptochilichthyidae), Argentinidae, Bathylaconidae, Microstomatidae (Bathylagidae), Opisthoproctidae, Platytroctidae] Salmoniformes [Coregonidae, Galaxiidae (Aplochitonidae, Lepidogalaxiidae), Osmeridae (Plecoglossidae, Salangidae), Retropinnidae (Prototroctidae), Salmonidae (Thymallidae)] Esociformes [Esocidae, Umbridae] Stomiiformes [Gonostomatidae, Photichthyidae, Sternoptychidae, Stomiidae (Astronesthidae, Idiacanthidae, Malacosteidae, Melanostomiidae)] Ateleopodiformes [Ateleopodidae] Aulopiformes [Alepisauridae (Omosudidae), Aulopidae, Bathysauridae, Chlorophthalmidae, Evermannellidae, Giganturidae, Ipnopidae (Bathypterodidae), Notosudidae (Scopelosauridae), Paralepididae (Anotopteridae), Pseudotrichonotidae, Scopelarchidae, Synodontidae (Harpadontidae)] Myctophiformes [Myctophidae, Neoscopelidae] Lampridiformes [Veliferidae, Lamprididae, Stylephoridae, Lophotidae, Radiicephalidae, Trachipteridae, Regalecidae] Polymixiiformes [Polymixiidae] Percopsiformes [Amblyopsidae, Aphredoderidae, Percopsidae] Ophidiiformes [Aphyonidae, Bythitidae, Carapidae (Fierasferidae), Ophidiidae (Brotulidae), Parabrotulidae] Gadiformes [Bregmacerotidae, Euclichthyidae, Gadidae (Gaidropsaridae, Lotidae, Phycidae, Ranicipitidae), Macrouridae (Bathygadidae, Trachyrincidae), Melanonidae, Merlucciidae (Macruronidae, Steindachneriidae), Moridae (Eretmophoridae), Muraenolepididae] Batrachoidiformes [Batrachoididae] Phylogeny and Systematics 23

Table 2.3 Continued

Lophiiformes [Antennariidae, Brachionichthyidae, Caulophrynidae, Centrophrynidae, Ceratiidae, Chaunacidae, Diceratiidae, Gigantactinidae, Himantolophidae, Linophrynidae, Lophiidae, Melanocetidae, Neoceratiidae, Ogcocephalidae, Oneirodidae, Thaumatichthyidae] Stephanoberyciformes [Barbourisiidae, Cetomimidae, Gibberichthyidae, Hispidoberycidae, Megalomycteridae, Melamphaidae, Mirapinnidae, Rondeletiidae, Stephanoberycidae] Zeiformes [Caproidae, Grammicolepididae, Macrurocyttidae, Oreosomatidae, Parazenidae, Zeidae] Beryciiformes (Trachichthyiformes) [Anomalopidae, Anoplogastridae, Berycidae, Diretmidae, Holocentridae, Monocentridae, Trachichthyidae] Mugiliformes [Mugilidae] Atheriniformes [Atherinidae, Atherionidae, Atherinopsidae, Melanotaeniidae (Bedotiidae, Pseudomugilidae, Telmatherinidae), Notocheiridae (Isonidae), Phallostethidae (Dentatherinidae, Neostethidae)] Beloniformes [Adrianichthyidae, Belonidae, Exocoetidae, Hemiramphidae, Scomberesocidae] Cyprinodontiformes [Anablepidae, Aplocheilidae, Cyprinodontidae, Fundulidae, Goodeidae, Poeciliidae, Profundulidae, Rivulidae, Valenciidae] Synbranchiformes [Chaudhuriidae, Mastacembelidae, Synbranchidae] Gasterosteiformes (Indostomiformes, Pegasiformes, Syngnathiformes) [Aulorhynchidae, Aulostomidae, Centriscidae, Elassomatidae, Fistulariidae, Gasterosteidae, Hypoptychidae, Indostomidae, Macroramphosidae, Pegasidae, Solenostomidae, Syngnathidae] Dactylopteriformes [Dactylopteridae] Gobiesociformes [Gobiesocidae (Alabetidae, Cheilobranchidae), Callionymidae, Draconettidae] Perciformes (Scorpaeniformes, Zoarciformes) [Abyssocottidae, Acanthuridae, Acropomatidae, Agonidae, Amarsipidae, Ammodytidae, Anabantidae, Anarhichadidae, Anoplomatidae, Apistidae, Aploactinidae, Aplodactylidae, Apogonidae, Ariommatidae, Arripidae, Badidae, Banjosidae, Bathyclupeidae, Bathydraconidae, Bathylutichthyidae, Bathymasteridae, Belontiidae (Luciocephalidae, Osphronemidae, Polyacanthidae), Bembridae, Blenniidae, Bovichtidae (Bovichthyidae), Bramidae, Butidae, Caesioscorpididae, Callanthiidae, Caracanthidae, Carangidae, Caristiidae, Centracanthidae, Centrarchidae, Centrogeniidae, Centrolophidae, Centropomidae, Cepolidae (Owstoniidae), Chaenopsidae, Chaetodontidae, Champsodontidae, Chandidae (Ambassidae), Channichthyidae (Chaenichthyidae), Channidae, Cheilodactylidae, Cheimarrhichthyidae, Chiasmodontidae, Chironemidae, Cichlidae, Cirrhitidae, Clinidae (Ophiclinidae, Peronedysidae), Comephoridae, Congiopodidae, Coryphaenidae, Cottidae, Creediidae (Limnichthyidae), Cryptacanthodidae, Cyclopteridae, Dactyloscopidae, Dichistiidae (Coracinidae), Dinolestidae, Dinopercidae, Drepanidae, Echeneidae, Eleotrididae, Embiotocidae, Emmelichthyidae, Ephippidae (Chaetodipteridae, Platacidae, Rhinopreneidae), Epigonidae, Enoplosidae, Ereuniidae, Gempylidae, Gerreidae, Girellidae, Glaucosomatidae, Gnathanacanthidae, Gobiidae (Amblyopidae, Cerdalidae, Gobioididae, Gobionellidae, Gunnellichthyidae, Kraemeriidae, Microdesmidae, Oxudercidae, Oxymetapontidae, Periophthalmidae, Ptereleotridae, Schindleriidae, Trypauchenidae), Grammatidae, Haemulidae (Plectorhynchidae, Pomadasyidae), Harpagiferidae (Artedidraconidae), Helostomatidae, Hemitripteridae, Hexagrammidae, Hoplichthyidae, Icosteidae, Inermiidae, Kuhliidae, Kurtidae, Kyphosidae, Labracoglossidae, Labridae (Callyodontidae, Scaridae, Odacidae), Labrisomidae, Lactariidae, Latidae, Latridae, Leiognathidae, Leptobramidae, Leptoscopidae, Lethrinidae, Liparidae, Lobotidae (Coiidae, Datnioididae), Lutjanidae (Caesionidae), Luvaridae, Malacanthidae (Branchiostegidae, Latilidae), Menidae, Microcanthidae, Monodactylidae, Moronidae, Mullidae, Nandidae, Nematistiidae, Nemipteridae, Neosebastidae, Nomeidae, Normanichthyidae, Nototheniidae, Odontobutidae, Opistognathidae, Oplegnathidae, Ostracoberycidae, Parabembridae, Parascorpididae, Pataecidae, Pempheridae, Pentacerotidae, Percichthyidae (Gadopsidae, Nannopercidae, Perciliidae), Percidae, Percophidae, Peristediidae, Pholidae, Pholidichthyidae, Pinguipedidae (Parapercidae), Platycephalidae, Plectrogeniidae, Plesiopidae (Acanthoclinidae, Notograptidae), Polynemidae, Polyprionidae, Pomacanthidae, Pomacentridae, Pomatomidae, Priacanthidae, Pristeolepidae, Pseudochromidae (Anisochromidae, Congrogadidae, Pseudoplesiopidae), Psychrolutidae, Ptilichthyidae, Rachycentridae, Rhamphocottidae, Rhyacichthyidae, Scatophagidae, Sciaenidae, Scombridae, Scombrolabracidae, Scombropidae, Scorpaenidae (Pteroidae), Scorpididae, Scytalinidae, Sebastidae (Sebastolobidae), Serranidae (Anthiidae, Niphonidae, Grammistidae, Pseudogrammatidae), Setarchidae, Siganidae, Sillaginidae, Sinipercidae, Sparidae, Sphyraenidae, Stichaeidae, Stromateidae, Symphysanodontidae, Synanceiidae (Inimicidae, Choridactylidae, Minoidae), Terapontidae, Tetragonuridae, Tetrarogidae, Toxotidae, Trachinidae, Trichiuridae, Trichodontidae, Trichonotidae, Triglidae, Tripterygiidae, Uranoscopidae (Xenocephalidae), Xenisthmidae, Xiphiidae (Istiophoridae), Zanclidae, Zaniolepididae, Zaproridae, Zoarcidae] Pleuronectiformes [Achiropsettidae, Archiridae, Bothidae, Citharidae, Cynoglossidae, Paralichthyidae, Paralichthodidae, Pleuronectidae, Poecilopsettidae, Psettodidae, Rhombosoleidae, Samaridae, Scophthalmidae, Soleidae] Tetraodontiformes [Balistidae, Diodontidae, Molidae, Monacanthidae, Ostraciidae (Aracanidae, Ostraciontidae), Triacanthidae, Triacanthodidae, Tetraodontidae (Canthigasteridae), Triodontidae] 24 Chapter 2 and rays). They are readily distinguished from benthic shark in one family, which live in coastal other fishes in having a cartilaginous skeleton waters less than 300m deep. They are readily dis- with distinctive superficial calcification, called tinguished from all other extant sharks in having, ‘prismatic’ calcification, and males with pelvic in combination, an anal fin and spines present at claspers, which are inserted into the female cloaca the front of each dorsal fin. and oviduct during copulation. Synapomorphies The Orectolobiformes (carpet sharks and allies) diagnosing the group are listed by Didier (1995, is a small but diverse order of about 30 species of table 1). mainly coastal, bottom-dwelling sharks. How- The Chimaeriformes (chimeras and allies) is ever, it also includes the pelagic, plankton-feeding a small order of about 50 or so species of mainly whale shark (Rhincodon typus), the world’s largest deepsea fishes that have a worldwide distribution. fish, which attains a length of at least 12m but They have soft, scaleless bodies with prominent can purportedly grow to 18m and weigh 12tonnes. sensory canals on the head, a prominent spine in Orectolobiforms are distinctive in having a short, front of the first dorsal fin, three pairs of beak-like broadly arched to nearly straight mouth that is en- teeth in the mouth, two in the upper jaw and one tirely in front of the eyes, and distinctively formed in the lower jaw, spiracle absent in adults, and a nasal flaps and barbels. Other characters defining fleshy operculum over the four gill openings. This the group and supporting internal relationships are means they have only one external opening. Males discussed by Dingerkus (1986), Compagno (1988), are distinctive in having, in addition to pelvic cla- Shirai (1996) and de Carvalho (1996). spers found in other chondrichthyans, spiny re- The Carcharhiniformes (requiem sharks and al- tractable appendages in front of the pelvic fins and lies) is a moderate-sized order of about 230 species on the head. These are used to clasp or stimulate of mainly marine sharks found throughout the the female during copulation. Classification and world. There are several species that enter or are relationships within the order are reviewed by confined to fresh water. The order includes many Didier (1995) based on an extensive array of mor- of the most familiar large predatory sharks, includ- phological characters. ing the tiger shark (Galeocerdo cuvier), whaler or All extant elasmobranchs belong to the requiem sharks (Carcharhinus) and hammerhead Neoselachii. Characters supporting monophyly sharks (Sphyrna and Eusphrya), as well as many of this group are discussed by de Carvalho (1996, smaller bottom or near-bottom dwelling sharks, p. 39). The composition and interrelationships of such as the catsharks (Scyliorhinidae) and hound- the orders of Neoselachii are controversial. The sharks (Triakidae). Carcharhiniforms are distin- classification and relationships presented herein guished from other sharks in having a nictitating follows de Carvalho (1996), who also reviews lower eyelid or fold, a transparent movable mem- previous hypotheses. The reader is referred to brane or inner eyelid that protects the eye during that work for details of character support feeding and helps keep it clean. Other characters and named supraordinal taxa. Thus, ten orders diagnosing the group and relationships within it are recognized: Heterodontiformes, Orectolobi- are discussed by Compagno (1988) and Shirai formes, Carcharhiniformes, Lamniformes, (1996). Hexanchiformes, Echinorhiniformes, Squali- The Lamniformes (mackerel sharks and allies) formes, Squatiniformes, Pristiophoriformes and is a small order of 16 species of coastal marine and Rajiformes. The orders are arranged into two major oceanic sharks. It is a very diverse group that in- groups by de Carvalho (1996; see also Shirai cludes, among others, the bizarre goblin shark 1992, 1996), Galeomorphii (= Galea of Shirai 1996) (Mitsukurina owstoni), the notorious great white for the first four orders, and Squalea for the remain- shark (Carcharodon carcharias), the filter-feeding ing six orders. basking shark (Cetorhinus maximus, the world’s The Heterodontiformes (hornsharks or Port second largest fish, attaining sizes of at least 10m), Jackson sharks) is a small order of nine species of the unusual thresher sharks (Alopidae) and the Phylogeny and Systematics 25

filter-feeding megamouth shark (Megachasma high-quality oil, squalene, which is stored in the pelagios). Considering that the latter grows to liver. Characters supporting monophyly and inter- lengths of over 5m, it is remarkable that the first relationships of the order are discussed by de specimen was not collected until 1976. Lamni- Carvalho (1996). forms are distinguished from other sharks in The Squatiniformes (angelsharks) is an order of having a distinctive tooth pattern, and uterine about 15 species of benthic sharks found in tropical cannibalism. Other characters diagnosing the to temperate seas at depths ranging from a few me- group are discussed by Compagno (1988). Rela- tres to 1300m. They are flattened ambush preda- tionships among lamniform families are contro- tors with laterally placed gill slits, pectoral fins versial (cf. Maisey 1985; Compagno 1990; Shirai that are free from the head, and the lower caudal 1996; de Carvalho 1996). lobe longer than the upper. They are commercially The Hexanchiformes (frill shark and cow fished for their flesh, oil and leather (Last and sharks) is a small order of five species of mainly Stevens 1994). deepsea sharks found in cold waters throughout The Pristiophoriformes (saw sharks) is an order the world. They are distinguished from other with seven or so species of mainly marine temper- sharks by the presence of six or seven gill slits and a ate to tropical sharks found throughout the West single, spineless, posteriorly positioned dorsal fin. Pacific, the southwestern Indian Ocean and the Other characters supporting monophyly of the tropical West Atlantic. They are distinctive fishes, group are discussed by de Carvalho (1996, p. 46). though they superficially resemble the rajiform The remaining five squalean orders form a family Pristidae (sawfishes), with a shark-like monophyletic group, which is diagnosed by, body, an elongate snout with enlarged tooth-like among other characters, the absence of an anal fin denticles on each side and a pair of barbels on the (de Carvalho 1996). The Echinorhiniformes (bram- underside, and four to six pairs of gill slits placed ble sharks) is a small order of two species of large laterally on the head. (2.6–4m), bottom or near-bottom dwelling sharks The Rajiformes (= Batoidea; rays and relatives) found in deep water throughout tropical and tem- is a moderately large order with over 450 species of perate seas. They are characterized by the presence mainly benthic marine fishes found worldwide. of two relatively small, spineless dorsal fins posi- There are also some freshwater species. It is a di- tioned well behind the pelvic origin, and coarse verse order and many workers advocate dividing it denticles or enlarged thorns on the skin. Addi- into four or more orders. It includes various bot- tional characters supporting monophyly of the tom or near-bottom species such as the shark-like order are discussed by de Carvalho (1996). rhynids (shark rays) and pristids (sawfishes), disc- The Squaliformes (dogfishes and relatives) is an shaped rajids (skates), and the electric torpedinids, order of about 70 species of mainly bottom or hypnids, narcinids and narkids (collectively called near-bottom living sharks found in all the world’s electric rays), as well as the huge pelagic Mobula oceans. Although some occur in relatively shallow and Manta (devil and manta rays; up to 6m across water, most species live in deep water down to and weighing over 1300kg). Members of the order depths exceeding 6km. They vary considerably in are distinguished from other elasmobranchs in morphology, ranging from species 16cm long to having ventrally positioned gill slits. Characters the Greenland shark (Somniosus microcephalus), supporting monophyly of the order (as the super- which at 6.4m is one of the largest sharks. The order Batoidea) are discussed by McEachran et al. bizarre, 45-cm long, pelagic cookie-cutter shark (1996). The familial classification of the order is (Isistius brasiliensis), which feeds by removing a controversial; the one presented here largely plug of flesh from larger fishes and cetaceans, is follows McEachran et al. (1996), who also review also in the order. Many of the deepwater species are relationships within the order. bioluminescent. Some squaliforms are important The remaining vertebrates are classified in the fisheries species, harvested either for food or for a Osteichthyes, which includes two extant groups, 26 Chapter 2 the Sarcopterygii and Actinopterygii. Characters and evolution of lungfishes is given in Bemis et al. diagnosing the Osteichthyes include the presence (1987). of ossified dermal opercular plate(s) covering The order Coelacanthiformes (coelacanths) has the gills laterally, branchiostegal rays, and an just two extant species of deepwater reef fishes, interhyal bone (Lauder and Liem 1983). Some one found in the western Indian Ocean off the authors (such as J.S. Nelson 1994) use the term coast of East Africa, and the other off northern ‘Teleostomi’ for the group we here call Oste- Sulawesi, Indonesia. They are remarkable for ichthyes, instead using the latter as a term for a the recency of their discovery, highlighting our phenetic grouping that includes only the bony poor understanding of deep-reef fishes in the fishes. Indo-Pacific: the African species (Latimeria The Sarcopterygii includes three extant fish chalumnae) was discovered in 1938, whereas orders (Ceratodontiformes, Lepidosireniformes the Indonesian species (L. menadoensis) was dis- and Coelacanthiformes) and the tetrapods. covered 60 years later in 1998 (Erdmann et al. Monophyly of the Sarcopterygii is supported by 1998). Forey (1998) provides a monograph on the numerous synapomorphies, particularly asso- systematics, evolution and biology of living and ciated with the structure of the skull and pectoral fossil coelacanthiforms. girdle (Cloutier and Ahlberg 1996). Relationships The Actinopterygii (ray-finned fishes) is a huge of the numerous fossil sarcopterygian fish taxa to diverse clade of fishes defined mainly on the basis each other and to the tetrapods is still in a state of of synapomorphies involving the morphology of flux, resulting particularly from an imprecise but the pelvic and pectoral girdles, scales, and median improving knowledge of fossils and their charac- fin rays, although most are secondarily lost or ters (Cloutier and Ahlberg 1996; Ahlberg and modified in higher actinopterygians. These char- Johanson 1998). However, there is consensus acters are reviewed by Lauder and Liem (1983). regarding the positions of the living sarcopterygian There has been considerable debate regarding fish orders relative to each other and to the the interrelationships of the basal actinoptery- tetrapods (Fig. 2.2): the Ceratodontiformes and gian orders Polypteriformes, Acipenseriformes, Lepidosireniformes belong to a monophyletic Amiiformes and Lepisosteiformes, and their Dipnoi (lungfishes), which is more closely related relationships to the Teleostei. For example, the to the Tetrapoda than it is to the final extant Holostei, a taxon based on the Amiiformes and sarcopterygian order, Coelacanthiformes Lepisosteiformes and various fossil taxa, has been (Cloutier and Ahlberg 1996). recognized historically. Gardiner et al. (1996) pre- Monophyly of the Dipnoi, a group best known sented morphological evidence to conclude that it for its 270 or so fossil species, is based mainly is paraphyletic, forming a series of sister groups to on synapomorphies associated with dentition and the Teleostei as indicated in Fig. 2.2. Conversely, skull morphology (reviewed by Cloutier and however, they also presented molecular evidence Ahlberg 1996). The Ceratodontiformes (Queens- for a monophyletic Holostei. Discussions con- land lungfish) includes a single extant species cerning the relationships of basal actinopterygians from fresh waters of southeastern Queensland, and supporting character evidence are provided by Australia. The Lepidosireniformes (South Grande and Bemis (1996, 1998) and Gardiner et al. American and African lungfishes) includes five ex- (1996). tant freshwater species, one from South America The order Polypteriformes (= Cladistia; bichirs and four from Africa. Unlike the Queensland lung- or featherfins) consists of about 10 species in a fish, the South American and African lungfishes single family of small freshwater fishes from West are able to tolerate environmental desiccation by Africa, although fossils are also known from North aestivating, either in a moist burrow (South Amer- Africa and possibly Bolivia. They have an unusual ican lungfish) or a dry cocoon (African lungfishes). dorsal fin consisting of a number of finlets. These A collection of papers on the biology, systematics are composed of a single spine to which one or Phylogeny and Systematics 27 more soft rays are attached, leading to the common evolution and biogeography of fossil and extant name ‘featherfin’. Other characters diagnosing the amiids is provided by Grande and Bemis (1998), order are summarized by J.S. Nelson (1994). Some who also investigate relationships among ami- workers have considered the order to belong in the iforms and other halecomorphs. Sarcopterygii, particularly based on their pectoral- All remaining fishes belong to a huge taxon fin structure, but current consensus favours a basal called the Teleostei. For most people, teleosts are position within the Actinopterygii (Gardiner and the typical fishes. Characters supporting mono- Schaeffer 1989). phyly of the Teleostei are reviewed by de Pinna The order Acipenseriformes (sturgeons and (1996). There are two competing hypotheses of re- paddlefishes) includes 27 species in two families lationships among basal extant teleosts. In one (see of Northern Hemisphere freshwater and anadro- Patterson 1998; Fig. 2.2), the Osteoglossomorpha mous fishes. They include some of the largest of is the sister group of a clade consisting of all fishes, with several species purportedly reaching other teleosts, of which the Elopomorpha is the lengths in excess of 7m, though confirmed records basalmost clade. In the other hypothesis (see are around 4–5m. Many acipenseriforms are, or Arratia 1998), the phylogenetic positions of the have been, commercially important, harvested Elopomorpha and Osteoglossomorpha are re- either for food (particularly their caviar), or for versed relative to other teleosts. their swimbladders, which yield isinglass, used for The Osteoglossomorpha is a small yet diverse making special glues and historically for clarifying group of freshwater fishes, characterized by the white wines. Well over half the acipenseriform primary bite being between the parasphenoid bone species are now considered endangered. Charac- on the lower part of the skull and the tongue. ters supporting the monophyly and interrelation- Monophyly of the group and interrelationships of ships of the Acipenseriformes, both fossil and included taxa both extant and fossil are reviewed recent, are reviewed by Grande and Bemis (1996). by Li and Wilson (1996). Two extant orders Diversity, biology and conservation of the order are currently recognized: the Hiodontiformes are reviewed by Birstein et al. (1997) and Secor et al. (mooneyes), a small order with two extant species (2000). from fresh waters of North America, although The Lepisosteiformes (= Ginglymodi; gars) is a fossil forms are also known from China; and small order (seven species in one family) of large the Osteoglossiformes (bonytongues, Old World freshwater and estuarine but rarely marine fishes knifefishes and elephantfishes), a moderately from eastern North America, Central America and small order with over 200 species from tropical Cuba. Fossil species are also known from South fresh waters of South America, Africa, Southeast America, West Africa, Europe and India. They are Asia, Australia and New Guinea. Fossil osteoglos- heavily scaled, predatory fishes with elongate siforms are also known from North America. The snouts; unlike other fishes with elongate snouts, Osteoglossiformes includes the South American such as belonids and sturgeons, the nostrils are arapaima (Arapaima gigas), one of the world’s located near the tip of the snout rather than near largest freshwater fishes, as well as the elephant- the eyes. Characters supporting monophyly of the fishes (Mormyridae), an African family of weakly Lepisosteiformes and relationships among its in- electric fishes. cluded species, both fossil and extant, are provided The Elopomorpha is a diverse group of mainly by Wiley (1976). marine fishes that are characterized by an unusual The Amiiformes (bowfins and allies) includes a leaf-like larval stage (leptocephalus) and a distinc- single extant species from fresh waters of eastern tive sperm morphology. Molecular and morpho- North America. However, fossil amiiforms are logical characters supporting monophyly of the known from freshwater and marine deposits in Elopomorpha and the composition and interrela- North America, South America, Eurasia and tionships of the four included orders are discussed Africa. A treatise on the systematics, anatomy, by Forey et al. (1996). About eight species are 28 Chapter 2 included in the Elopiformes (tarpons and allies), Johnson and Patterson (1996) in the composition which are mainly marine herring-like fishes, al- of the Protacanthopterygii (Salmoniformes and though some enter fresh water. They are found Argentiniformes; see below), and in recognizing a throughout tropical and subtropical regions. sister relationship between the Esociformes and The tarpon (Megalops atlanticus) is an important Neoteleostei. The latter relationship follows gamefish. The Albuliformes (bonefish and allies) Parenti (1986) and is based on the mode of tooth includes about six species of mainly marine fishes attachment and the presence of acellular bone. found throughout tropical and subtropical regions. The Clupeomorpha includes a single extant About 25 species of elongate deepsea fishes make order, Clupeiformes (herrings, anchovies and up the Notacanthiformes (halosaurs and spiny relatives), of freshwater and coastal marine fishes eels). The final order, Anguilliformes (eels and found globally in temperate to tropical areas. gulpers), includes over 700 species of mainly ma- There are about 360 or so extant species included rine (some freshwater) elongate fishes. Relation- in the order, among them some of the world’s most ships and classification within the Anguilliformes important commercial species. They are typically are controversial. For example, the deepsea gulpers silvery fishes with a pelvic scute, which is and allies (suborder Saccopharyngoidei) and the sometimes inconspicuous, usually a median row eels (suborder Anguiloidei) are sometimes classi- of scutes (often few in number or absent) along the fied as separate elopomorph orders, but evidence abdomen in front of and behind the pelvic fins, and presented by Forey et al. (1996) suggests that primitively with an additional median row of some anguiloids are more closely related to sac- scutes in front of the dorsal fin. Anatomically they copharyngoids than they are to other anguilloids are characterized by a distinctive connection be- (however, see Robins 1997). The familial classifi- tween the ear and the swimbladder. Additional cation of the Anguilliformes has not been cladisti- characters supporting monophyly of the group cally demonstrated. are discussed by Lecointre and Nelson (1996), The remaining teleosts, collectively termed the who also summarize internal relationships. The Clupeocephala, are diagnosed by synapomorphies neotenic Southeast Asian freshwater genus Sun- associated with the lower jaw and caudal skeleton dasalanx was recently placed with the clupeiform (Patterson and Rosen 1977, p. 130). The following family Clupeidae by Siebert (1997). Previously taxa are usually recognized in the Clupeocephala: it had been assigned to its own family in the Clupeomorpha, Ostariophysi, Protacanthoptery- Salmoniformes (Roberts 1981). We have adopted a gii, Esociformes and Neoteleostei. Relationships conservative, though cladistically defensible, fa- among these taxa are controversial. Thus, the milial classification largely following J.S. Nelson relationships presented (Fig. 2.2) are likely to be (1994) (Table 2.3). Other authors have recognized unstable. We follow Lecointre and Nelson (1996) additional families (see J.S. Nelson 1994 for in recognizing a sister relationship between the review). Clupeomorpha and the Ostariophysi. This rela- The Ostariophysi is a huge group of mainly tionship is mainly supported by molecular charac- freshwater teleosts that includes nearly 75% of all ters, although these have not been extensively freshwater fishes; new species continue to be dis- surveyed in basal clupeocephalans. Morphological covered at a rapid rate, particularly in Southeast evidence possibly bearing on the relationship, Asia and the Neotropics (Lundberg et al. 2000). Os- such as fusion of various bones in the caudal skele- tariophysans are distinctive in possessing epider- ton and morphology of the sensory canal in the mal alarm substance cells and an alarm reaction. A skull, is ambiguous. An alternative to this rela- wounded fish releases alarm substance into the tionship is one in which the Clupeomorpha is the surrounding water, which triggers alarm reactions sister group of a clade, called the Euteleostei, made such as scattering behaviour in adjacent fishes. up of ostariophysans and the remaining clupeo- Interestingly, alarm substance and alarm reaction cephalans (Patterson and Rosen 1977). We follow are lacking in the electrocommunicating gymnoti- Phylogeny and Systematics 29 form ostariophysans. Other characters supporting species of compressed or cylindrical eel-like monophyly of the Ostariophysi and relationships Neotropical fishes make up the Gymnotiformes among the five included orders are reviewed and (American or electric knifefishes), although it is discussed by Fink and Fink (1996). The interordi- sometimes classified as a suborder of the nal relationships summarized in Fig. 2.2 fol- Siluriformes (e.g. Fink and Fink 1981). These fishes low Fink and Fink. The Gonorynchiformes (= are remarkable in possessing an organ system for Anotophysi; milkfish and allies) is a small order of producing and receiving electrical impulses; the about 35 species of Indo-Pacific marine and electric eel (Electrophorus) is a famous, though African freshwater fishes. Reviews of relation- atypical, example. A phylogeny and classification ships within the Gonorynchiformes are provided of the order, based on morphological and behav- by Fink and Fink (1996) and Johnson and Patterson ioural characters, including some from electric (1997). The remaining ostariophysan orders are organ morphology and discharge patterns, is pre- grouped into a taxon called the Otophysi. This is sented by Albert and Campos-da-Paz (1998). one of the best circumscribed higher fish taxa, The Protacanthopterygii includes two orders, which is defined in part by a complex of modifica- Argentiniformes and Salmoniformes. There has tions of the anterior vertebrae and associated been considerable debate over the composition of elements that are collectively called a Weberian these two orders and relationships of the fishes apparatus. The carps, loaches and suckers belong included in them (reviewed by Johnson and to the Cypriniformes, a large order of about 2700 Patterson 1996). Particular controversy has cen- species of freshwater and rarely brackish water tred around the positions of the Northern fishes found throughout North America, Africa Hemisphere osmeroids, the Southern Hemisphere and Eurasia. Interrelationships of the five cyprini- galaxioids and the enigmatic Western Australian form families are reviewed by Fink and Fink salamanderfish (Lepidogalaxias). The Protacan- (1996). The Cyprinidae is worthy of mention, as it thopterygii as defined by Greenwood et al. (1966) is one of the largest of all fish families (with about was a much larger, poorly defined group that in- 2000 species), and includes such familiar species as cluded, in addition to the above, taxa now assigned Brachydanio rerio, the zebra danio or zebrafish, an to the Gonorynchiformes, Esociformes, Stomii- important subject for genetic and development formes, Ateleopodiformes, Aulopiformes, Mycto- studies, Carassius auratus (the goldfish) and phiformes and Stephanoberyciformes; some of Cyprinus carpio (the carp); the latter two species these were placed within the Salmoniformes. Even are remarkable in that they have been introduced as currently recognized, character support for widely throughout the world’s fresh waters. The the Protacanthopterygii is relatively weak, being Characiformes (characins, tetras and allies) is a based on two non-unique characters associated large order with around 1400 species of Neotropi- with the structure of intermuscular bones cal and African freshwater fishes. Relationships (Johnson and Patterson 1996, p. 314). and familial classification of characiforms are con- The Salmoniformes, containing the trouts, troversial (see review by Vari 1998); the familial salmons, smelts and relatives, is a small order of classification presented here (Table 2.3) follows about 175 species of marine, freshwater, anadro- Buckup (1998). The Siluriformes (catfishes) is a mous and diadromous fishes. Kottelat (1997) pro- large order of about 2500 species of mainly fresh- vided justification for the recognition of many water and some marine fishes found throughout additional species of European coregonids and the tropics and some temperate areas. The familial salmonids. They include some of the most impor- classification and interrelationships of siluriforms tant commercial and recreational fisheries species are controversial; the classification in Table 2.3 in the world; Salvelinus fontanalis (brook trout), follows Ferraris and de Pinna (1999), although this Salmo trutta (brown trout) and Onchorhynchus latter work does not exactly match the cladistic mykiss (rainbow trout) have been introduced scheme outlined by de Pinna (1998). About 150 almost the world over for these purposes. 30 Chapter 2

Salmoniforms are distinguished by various organs, and these have a distinctive structure that synapomorphies, including scales without radii distinguishes stomiiforms from all other fishes and absence of epipleural bones. Additional char- with light organs. Other characters supporting acters supporting monophyly and relationships monophyly of the order, as well as relationships within the order are reviewed by Johnson and within it, are reviewed by Harold and Weitzman Patterson (1996). (1996). The Argentiniformes (deepsea smelts, slick- About 12 species of bottom-dwelling deepsea heads and relatives) is a small order of oceanic fishes comprise the Ateleopodiformes (snotnose or and deepsea fishes found throughout the world’s jellynose fishes). They are found worldwide, and oceans, and contains about 170 species. They are are elongate flabby fishes with a weakly ossified distinguished by the presence of a crumenal organ, skeleton, a long anal fin continuous with the cau- a unique bilaterally paired pouch-like structure dal, a short dorsal fin positioned near the head, an derived from the upper part of the posterior two inferior mouth and a pronounced jelly-like snout. gill arches and the front part of the oesophagus The largest species grow to nearly 2m. They have (Greenwood and Rosen 1971). Other characters been classified near the lampridiform fishes, but supporting monophyly and relationships within Olney et al. (1993) removed them from the the order are reviewed by Johnson and Patterson Acanthomorpha. (1996). The Eurypterygii was diagnosed by Johnson The Esociformes is a small order of about 12 (1992) on the basis of several synapomorphies species of Northern Hemisphere freshwater associated with the gill arches, such as the tooth- fishes. It includes just two families, the Esocidae plate fused to epibranchial 3, presence of an (pikes) and Umbridae (mudminnows), and is interoperculoid ligament in the jaws, and fusion characterized by posteriorly positioned dorsal and of the ventral hemitrich of the medial pelvic ray anal fins. Characters supporting monophyly of to the medial pelvic radial. It includes two groups, the order are reviewed by Johnson and Patterson the Aulopiformes (= Cyclosquamata) and the (1996). The phylogenetic position of the Esoci- Ctenosquamata. formes is controversial, and it is often included in There are about 220 species of marine benthic the Salmoniformes. and pelagic fishes in the Aulopiformes (lizard- The remaining fish taxa are placed within the fishes and relatives). They occur worldwide in Neoteleostei. Characters supporting monophyly habitats ranging from coastal areas and estuaries to of this huge taxon are discussed by Johnson (1992). the abyss. The diversity of their habits is matched It can be divided into three taxa, Stomiiformes, in their morphology, ranging from small (10cm) Ateleopodiformes and Eurypterygii. Relation- sand-diving pseudotrichonotids to large (2m) ships among these three taxa are unresolved pelagic alepisaurids (lancetfishes). Despite this (Olney et al. 1993; Fig. 2.2). diversity, the monophyly of the order is well The Stomiiformes (dragonfishes and relatives) supported by various synapomorphies, mainly includes over 320 species of midwater and deep- associated with gill-arch morphology. Characters water oceanic fishes found worldwide. It is a very supporting aulopiform monophyly, intrarelation- diverse group, ranging from darkly pigmented, ships and familial classification are detailed by slender-bodied, large-fanged fishes, such as the Baldwin and Johnson (1996). Historically, aulopi- dragonfishes (Stomiidae), to the silvery, deep- forms have been incorrectly classified with the bodied hatchetfishes (Sternoptychidae). The group Myctophiformes. also includes tiny fishes of the genus Cyclothone, The myctophiforms and acanthomorphs, the delicate plankton dwellers, and considered by remaining teleosts, are grouped into a large taxon some scientists to be the most abundant vertebrate called the Ctenosquamata. Characters supporting genus in the world (see Miya and Nishida 1996). monophyly of this group are reviewed by Stiassny All but a single species of the order have light (1996). Phylogeny and Systematics 31

The Myctophiformes (= Scopelomorpha; vertebra, the Atherinomorpha (Atheriniformes, lanternfishes and relatives) is a moderate-sized Beloniformes and Cyprinodontiformes), Gas- order of about 250 species of small, mainly terosteiformes, Mastecembeloidei and Mugilidae mesopelagic fishes found in oceanic and coastal (which are both usually classified in the waters worldwide. Characters supporting mono- Perciformes), Synbranchiformes, and the genus phyly and interrelationships of the order are dis- Elassoma, which was previously classified in the cussed by Stiassny (1996). North American perciform family Centrarchidae, The Acanthomorpha is an immense grouping of form a monophyletic group, the Smegmamorpha. fishes that includes about 60% of all fish species, Johnson and Patterson’s Smegmamorpha has not and thus nearly one-third of all vertebrates, or been cladistically tested, but Parenti and Song about 14700 species. It is diagnosed by a number (1996) noted that some smegmamorphs (Elassoma of osteological features, the most conspicuous and mugilids) share a unique derived pattern of of which is the presence of true fin spines pelvic innervation with higher acanthomorphs (unsegmented, bilaterally fused anterior rays) in (such as Perciformes), but that atherinomorphs the dorsal and anal fins (Johnson and Patterson and gasterosteiforms possess a primitive arrange- 1993), though these are secondarily absent ment. This suggests that the Smegmamorpha is in some acanthomorphs. About 20 or so not monophyletic. In contrast, however, Johnson acanthomorph orders are usually recognized at and Springer (1997) have concluded that Elassoma present. We recognize 21 here: Lampridiformes, is a gasterosteiform. Polymixiiformes, Percopsiformes, Ophidii- Johnson and Patterson (1993) also proposed that formes, Gadiformes, Batrachoidiformes, Lophii- the Beryciformes of previous authors represents formes, Stephanoberyciformes, Zeiformes, two distinct lineages, the Stephanoberyciformes Beryciformes, Mugiliformes, Atheriniformes, and the Beryciformes. Moore (1993) agreed that Beloniformes, Cyprinodontiformes, Synbranchi- the traditional Beryciformes is not monophyletic, formes, Gasterosteiformes, Dactylopteriformes, but proposed a radically different arrangement in Gobiesociformes, Perciformes, Pleuronecti- which the vast majority of the traditional taxa, formes and Tetraodontiformes. including Johnson and Patterson’s Stephanobery- Relationships among acanthomorphs are ciformes, were classified in a separate order, controversial and represent one of the major the Trachichthyiformes, which is not recognized challenges facing vertebrate systematics. As G.J. here. Nelson (1989b, p. 328) put it when commenting on It is likely that debate over the relationships of our understanding of relationships among teleost acanthomorphs will continue. In addition to the fishes: ‘recent work has resolved the bush at the need for testing the relationships and characters bottom [relationships among basal teleosts], but proposed by Johnson and Patterson (1993), there is the bush at the top [the acanthomorphs] persists’. a need to tackle major questions not addressed by The most recent comprehensive treatment of these authors. In particular, the monophyly and acanthomorphs was by Johnson and Patterson intrarelationships of the Paracanthopterygii (1993), who on the basis of a large array of morpho- and Perciformes (see below), and the influence of logical characters, particularly those associated their probable non-monophyly on Johnson and with the configuration of the ribs and associated Patterson’s scheme of relationships, represent intermuscular bones, suggested a novel scheme of significant challenges for systematic ichthyology relationships among acanthomorphs. Johnson and (Gill 1996). Patterson’s phylogeny is largely followed here (Fig. The oarfishes and allies form a small order of 2.2). Among their major conclusions were that, about 20 species of oceanic fishes called the based on mainly osteological characters, such as Lampridiformes. Although very different in biolo- the first intermuscular bone (epineural) originat- gy and form, ranging from the deep-bodied lampri- ing at the tip of the transverse process on the first dids (opah) to the very elongate regalecids 32 Chapter 2

(oarfishes), lampridiforms are characterized by a have concluded that the order is not monophy- unique mouth structure. Other characters sup- letic. For example, Rosen (1985; see also Patterson porting monophyly of the order and relationships and Rosen 1989) was unable to find characters to among the seven families included are discussed diagnose the order but provided synapomorphies, by Olney et al. (1993). The oarfish (Regalecus such as a thoracic anus and segmented premaxilla, glesne) is the longest bony fish, purportedly reach- to link the families Aphredoderidae (pirate ing lengths of up to 17m, although confirmed re- perches) and Amblyopsidae (cavefishes) into a ports are to only 8m. It is almost certainly monophyletic group, the Aphredoderoidei. More responsible for early accounts of sea serpents, par- recently, Murray and Wilson (1999) proposed a dif- ticularly those that refer to monsters having ‘the ferent arrangement in which aphredoderids were head of a horse with a flaming red mane’ (Norman grouped with the Percopsidae (troutperches) and and Fraser 1937, p. 113). several fossil taxa in the Percopsiformes, and The Polymixiiformes (beardfishes) is a small amblyopsids were placed in their own order order with 10 species in one family, the Polymixi- (Amblyopsiformes) within the Anacanthini, a idae (Kotlyar 1996), and occurs on outer continen- group that also included gadiforms, ophidiiforms, tal slopes in tropical and subtropical waters of all lophiiforms and batrachoidiforms. major oceans. They are distinctive in having a pair The Ophidiiformes (cusk-eels, pearlfishes and of hyoid barbels, which leads to their common relatives) is a moderately large order with about name of beardfishes. Polymixiids have been vari- 370 species in five families of mainly marine ously classified in the Beryciformes (J.S. Nelson coastal to abyssal fishes, although there are a few 1984; Kotlyar 1996) or the superorder Paracan- brackish water or freshwater species (Nielsen et al. thopterygii (Rosen and Patterson 1969), but they 1999). They are characterized by pelvic fins, when are generally seen now as basal acanthomorphs present, anteriorly positioned beneath the head (Rosen 1985; Stiassny 1986; Johnson and Patterson and long-based dorsal and anal fins. However, 1993). monophyly of the order and its included families is Five acanthomorph orders, the Percopsiformes, questionable; for example, Patterson and Rosen Ophidiiformes, Gadiformes, Lophiiformes and (1989) concluded that the order is paraphyletic Batrachoidiformes, were grouped into a diverse with some ophidiiforms more closely related to a assemblage called the Paracanthopterygii by clade consisting of the Gadiformes, Lophiiformes Patterson and Rosen (1989), who also postulated and Batrachoidiformes than they are to other interordinal relationships. This superorder was ophidiiforms. Murray and Wilson (1999) also con- originally proposed by Greenwood et al. (1966), but cluded that the order is not monophyletic, but at that time also included several additional taxa. divided the group differently. An alternative view However, even as redefined by Patterson and of ophidiiform intrarelationships and classifica- Rosen and later supported by Murray and Wilson tion is discussed by Howes (1992). (1999), there is equivocal support for monophyly With about 500 species (Cohen et al. 1990), the of the Paracanthopterygii (Gill 1996) and ordinal Gadiformes (cods and relatives) is a moderately interrelationships and composition are con- large order. They are mainly marine fishes found in troversial (cf. Markle 1989; Patterson and Rosen temperate or deepsea habitats worldwide, and in- 1989; Murray and Wilson 1999). It is unlikely that clude the commercially important family Gadidae paracanthopterygian interrelationships will be and the large mainly benthopelagic family resolved without first resolving the question of Macrouridae. A very entertaining history of the monophyly of the group. Atlantic cod (Gadus morhua) fishery is provided The Percopsiformes (troutperches, pirate perch by Kurlansky (1997). Monophyly of the Gadi- and cavefishes and allies) is a small order of North formes is supported by various osteological and American freshwater fishes, with nine extant larval characters (Markle 1989), but relationships species in three families. However, some authors and familial classification within the order are Phylogeny and Systematics 33 controversial (e.g. see papers in Cohen 1989). Table whalefishes and allies) is an order of about 90 2.3 follows Cohen et al. (1990) with parenthetical species of small deepsea fishes found throughout families (synonyms) as recognized by Howes the world’s oceans. They are distinguished by a (1991), who provided systematic and distribu- number of osteological characters, such as the ab- tional summaries for each. sence of a subocular shelf on the infraorbital bones The Batrachoidiformes (toadfishes) includes and the presence of a uniquely hypertrophied one family and about 70 species of mainly marine extrascapular bone (Johnson and Patterson 1993). coastal fishes, with a few species confined to fresh Other characters supporting monophyly and water. The order has a worldwide distribution. As intrarelationships of the group are discussed by the scientific and common names suggest (batra- Moore (1993). Moore classified the taxon as one chos is Greek for frog), they resemble frogs and of two suborders in the Trachichthyiformes, the toads. They have smooth slimy skin with scales other suborder being the Trachichthyoidei; we absent or small and cycloid, prominent forward- follow Johnson and Patterson (1993), however, pointing eyes, ambushing predatory tactics, a large in including the latter in the Beryciformes. flattened head with a broad mouth, and the ability The Zeiformes (dories and allies) is an order to produce croaking or grunting sounds. Some with only about 40 species of deep-bodied fishes members of one subfamily (Porichthyinae) have found throughout the world’s oceans, mainly on rows of bioluminescent light organs, or photo- the continental slope and around seamounts, usu- phores, along the sides of their head and ally at depths of 600m or less. They have highly body. Members of another subfamily, the Thalas- protrusible jaws, and dorsal and anal fins with sophryninae, possess venomous spines on the spines anteriorly and unbranched segmented rays dorsal fin and gill cover. posteriorly. The soft and spinous portions of the The Lophiiformes (anglerfishes and relatives) is dorsal fin are separated by a notch. Monophyly of a moderately small, worldwide order with about the order is questionable; in particular, the inclu- 300 species of marine shallow to deepsea fishes. sion of the Caproidae is doubtful (see Johnson and Among the taxa included in the order are the Patterson 1993, p. 592). Rosen (1984) classified commercially important lophiids (monkfishes zeiforms within the Tetraodontiformes and pro- and goosefishes), the reef-dwelling antennariids posed that the Caproidae is the sister of a clade (frogfishes), the endangered southern Australian consisting of all other members of the redefined brachionichthyids (handfishes), and the deepsea Tetraodontiformes, which in turn is divisible into ceratioids (deepsea anglerfishes). The latter group two sister taxa, Zeomorphi (traditional Zeiformes are distinctive in showing extreme sexual dimor- minus Caproidae) and Plectognathi (traditional phism; in some, the males are very reduced in size Tetraodontiformes). and have denticular teeth for attaching to and There are about 150 species in the Beryciformes parasitizing females (Bertelsen 1984). Monophyly (squirrelfishes and allies). These are shallow to of the Lophiiformes is supported by various deepsea fishes found throughout the world’s synapomorphies; most notably the anterior dorsal- oceans. Monophyly of the order is supported by the fin ray, if present, is modified into an illicium (‘rod presence of Jakubowski’s organ, a uniquely inner- and line’) and esca (‘bait’), devices for attracting vated portion of the laterosensory system on the prey to the mouth. They also produce eggs snout (Johnson and Patterson 1993). However, ‘spawned embedded in a continuous ribbon-like the composition of the order is controversial (see sheath of gelatinous mucous’ (Pietsch 1984, Stephanoberyciformes above). Most members of p. 320). Additional characters supporting mono- the order avoid light; the shallow-water represen- phyly of the order and interrelationships among tatives, such as the squirrelfishes (Holocentridae) its included suborders are discussed by Pietsch and most flashlight fishes (Anomalopidae), are (1984) and Pietsch and Grobecker (1987). nocturnal, inhabiting caves or deeper waters The Stephanoberyciformes (pricklefishes, during the day. There are several commercially 34 Chapter 2 important species in the order, most notably the Parenti (1981), Collette et al. (1984) and Lovejoy orange roughy (Hoplostethus atlanticus). (2000). The Atheriniformes (silversides, rainbow- The Mugiliformes (grey mullets) is a small fishes and allies) is a moderately small order of order with one family containing about 80 species about 300 species of freshwater and coastal marine of coastal marine and freshwater fishes found fishes found throughout tropical to temperate throughout the world. They have two widely waters of the world. They are small fishes, usually separated dorsal fins, a small triangular mouth, with two dorsal fins with the anterior one having and pelvic fins positioned well behind the pectoral spines, a weak spine at the front of the anal fin, no fins. Many species are commercially important for lateral line and a silvery stripe along the side of the their flesh and roe. body, which gives the common name. There has The following three orders (Atheriniformes, Be- been considerable debate about the monophyly loniformes and Cyprinodontiformes) are grouped and relationships of atheriniform fishes in recent together in a taxon called the Atherinomorpha. years. This has particularly centred around the Monophyly of the Atherinomorpha is supported various taxa included here in the Melanotaeni- by a range of reproductive and osteological charac- idae (Bedotiidae, Pseudomugilidae and Telmath- ters (reviewed by Parenti 1993). They are remark- erinidae), Dentatherinidae and Phallostethidae, able for their diversity of reproductive modes. and their relationships to other atherinomorphs Although most are oviparous, viviparity has (e.g. Allen 1980; Parenti 1984; Stiassny 1990; Saeed evolved independently several times. Develop- et al. 1994; Dyer and Chernoff 1996; Aarn and ment is relatively slow, making atherinomorphs Ivantsoff 1997; Aarn et al. 1998). We tentatively ideal subjects for embryological and associated follow Dyer and Chernoff (1996) in accepting a studies. In certain species, such as the ‘annual monophyletic Atheriniformes. This is in contra- killifishes’ of the families Rivulidae and diction to Stiassny (1990), who proposed that the Aplocheilidae, eggs enter diapause during periods Bedotiidae is the sister of a clade consisting of of desiccation. The moderately large order Cyprin- all atherinomorphs. We also follow Dyer and odontiformes (killifishes and allies) contains Chernoff (1996) in the familial classification about 800 species found throughout freshwater presented (Table 2.3), but anticipate that debate and coastal marine areas of the New World, Africa, is likely to continue. southern Europe and Asia. Monophyly of the order The Synbranchiformes (swamp eels and allies) and interrelationships of its included genera were is a small order of about 90 species of freshwater investigated by Parenti (1981) and Costa (1998). and estuarine fishes found throughout the tropics. The Beloniformes (needlefishes and allies), at They are characterized by an eel-like body shape around 200 species, is a moderately small order and a distinctive configuration of the anterior of freshwater, coastal marine and oceanic fishes vertebrae (Johnson and Patterson 1993). Two found throughout tropical to temperate areas of suborders are currently included in the order. The the world. It is a relatively diverse order, with such Synbranchoidei (swamp eels) are found through- distinctive fishes as needlefishes (Belonidae), out the Neotropics, Africa, Asia and the Indo- slender predators with elongate jaws; flying fishes Australian area, and have highly reduced fins and a (Exocoetidae), pelagic fishes with enlarged pec- single ventral gill opening. Most are burrowers and toral and sometimes pelvic fins for gliding flight; many are amphibious, able to remain active out of and the ricefishes (Adrianichthyidae), small water for up to six months (Liem 1998). The Maste- killifish-like species from east Asia. Beloniforms cembeloidei (freshwater spiny eels) are found are distinctive in having more fin rays in the lower throughout Asia and Africa, and are also burrow- lobe of the caudal fin than the upper lobe, and the ing fishes, though most are fully aquatic. Until upper jaw fixed and non-protrusible. Additional recently, the Mastecembeloidei were classified as characters supporting monophyly of the order and a perciform suborder. relationships within it are discussed by Rosen and The Gasterosteiformes (pipefishes, stickle- Phylogeny and Systematics 35 backs and allies) is an order of about 280 species of Dactylopteridae) of bottom-living fishes, found freshwater and coastal marine fishes found throughout shallow to moderate-depth waters of throughout the world. It is a morphologically di- warm-temperate to tropical seas. They are distinc- verse group that includes some of the most bizarre tive in having a bony head, with a large con- fishes, such as seahorses, seadragons and pipe- spicuous preopercular spine, and enlarged pectoral fishes (Syngnathidae), ghost pipefishes (Solenosto- fins, with the upper five to seven rays short midae), shrimpfishes (Centriscidae), sea moths and separated from the remaining rays. Other (Pegasidae), paradox fishes (Indostomidae) and characters diagnosing the group are discussed by sticklebacks (Gasterosteidae). The latter family Gill (1890) and Eschmeyer (1997). Most recent includes some of the most intensively studied authors have classified the Dactylopteridae in the of all fishes, with many works devoted to their order Scorpaeniformes (here grouped with the biology, ethology and geographical variation (e.g. Perciformes), which they superficially resemble; Bell and Foster 1994). There is consensus neither dactylopterids bear, in particular, a striking on the composition of the order nor its internal re- resemblance to the family Triglidae (gurnards). lationships. Debate has in particular concentrated However, we follow Mooi and Gill (1995) in reject- on whether the sea moths and the paradox fishes ing a close relationship between ‘scorpaeniforms’ should be included in the order (see Johnson and and dactylopteriforms. Imamura (2000) has re- Patterson 1993). As noted above, the Elassomati- cently proposed that dactylopterids are nested dae, previously classified in the Perciformes, has within the perciform family Malacanthidae. also been recently added to the order by Johnson The Perciformes (perches and allies) is a huge and Springer (1997). order of fishes found throughout fresh and marine There are about 300 species in the Gobiesoci- waters of the world. With over 11000 species, it is formes (clingfishes, dragonets and allies), a small by far the largest vertebrate order; it is also one of order of mainly shallow marine fishes, though a the most diverse, ranging from tiny gobiids, in- few enter fresh water. They are found throughout cluding the 10-mm Trimmatom nanus, the short- tropical to temperate areas of the world, and are est vertebrate, to the 4.5-m pelagic black marlin distinctive in lacking scales and in having the pec- (Makaira indica), and, in particular, includes the toral and pelvic fins connected by a membrane and vast bulk of the diversity of coastal and reef fishes the infraorbital bone series represented only by a worldwide (Johnson and Gill 1998). However, such single anterior element, the ‘lachrymal’. Other claims are empty, as there is no evidence that characters diagnosing the order are summarized perciforms form a monophyletic group. Moreover, by Gill (1996). The three families comprising not only are there no synapomorphies to unite the the order have been variously classified in the order, it is not diagnosable even on the basis of Perciformes or Paracanthopterygii, and not always shared primitive characters. Thus the Perciformes associated with each other. For example, many has its basis in tradition rather than characters, and classifications place the gobiesocids in their own it is possible that the nearest relatives of some or order, Gobiesociformes, within or near the Para- many perciform taxa lie elsewhere in the Acantho- canthopterygii, while placing the draconnetids morpha. Relationships and classification within and callionymids in their own perciform suborder, the Perciformes are likewise dubious. Although Callionymoidei. Prior to Springer and Fraser some suborders (e.g. Acanthuroidei, Anaban- (1976), the relationships of the Australian gobieso- toidei, Blennioidei, Gobioidei, Notothenioidei, cid genus Alabes were particularly confused. On Scombroidei and Stromateoidei) and many the basis of its eel-like body form and single ventral families are demonstrably monophyletic, the gill opening, most previous authors had classified monophyly of many others, such as the suborders it in its own family in the Synbranchiformes. Labroidei and Tachinoidei, is dubious, and the The Dactylopteriformes (flying gurnards) is a Percoidei, which is the largest suborder with about small order (about seven species in a single family, 3000 species, is effectively a wastebasket for perci- 36 Chapter 2 form taxa that have not been assigned to other sub- Tetraodontiforms are often regarded as perciform orders (Gill and Mooi 1993; Johnson 1993; Mooi derivatives and, in particular, are often associated and Gill 1995; Mooi and Johnson 1997). We follow with the perciform suborder Acanthuroidei. Mooi and Gill (1995) in including the Scorpaeni- However, Rosen (1984) considered and rejected formes in the Perciformes; not only is there no such a relationship in favour of one between the justification for excluding scorpaeniforms, but Tetraodontiformes and Zeiformes (see above). they share a specialized arrangement of the epaxial musculature with certain traditional perciforms (including Perca, type genus of the order) that is 2.4 CONCLUSIONS not found in other Perciformes. Families listed in Table 2.3 follow Imamura and Shinohara (1998) Systematic ichthyology is an active and important and Johnson and Gill (1998) with some minor field. Our understanding of the relationships of updates. fish taxa and their associated classification contin- The Pleuronectiformes (flatfishes), with about ues to develop as new characters are investigated 570 species, is a moderately large order of mainly and as our understanding of character distribution marine fishes found worldwide. The order improves. Our appreciation of the diversity of contains many important commercial fishes, fishes also continues to develop, as new species are including halibuts, flounders and soles. described either through discovery of forms new to Pleuronectiforms are characterized by a unique science or through refinement of species concepts. synapomorphy: ontogenetic migration of one eye Other fields of fisheries biology, such as ecology, to the opposite side of the head. Adult flatfishes lie biogeography, genetics, development, physiology on their blind side, often with the body partially or and conservation, are recognizing that an under- completely covered with sand or silt; many are standing of the systematics of fishes can often be well known for their ability to rapidly alter their necessary for clarifying concepts and frequently coloration to match the substrate. Additional crucial to forming coherent and complete hy- characters supporting monophyly of the order and potheses. The rest of the Handbook will illustrate familial relationships within it are discussed by the significance of this. Chapleau (1993) and Cooper and Chapleau (1998). The Pleuronectiformes is sometimes stated as being derived from a perciform or percoid ancestor REFERENCES (e.g. J.S. Nelson 1994). However, there is no char- acter evidence for a relationship with or within Aarn and Ivantsoff, W. (1997) Descriptive anatomy perciforms. of Cairnsichthys rhombosomoides and Iriatherina The Tetraodontiformes is a moderately large werneri (Teleostei: Atheriniformes), and a phylo- order of about 340 species of mainly marine fishes genetic analysis of Melanotaeniidae. Ichthyological found throughout tropical to warm-temperate Exploration of Freshwaters 8, 107–50. Aarn, Ivanstoff, W. and Kottelat, M. (1998) Phylogenetic areas of the world. There are also species found in analysis of Telmatherinidae (Teleostei: Atherinomor- fresh water. It is a relatively diverse order, ranging pha), with description of Marosatherina, a new genus from the 2.5-cm diamond leatherjacket (Rudarius from Sulawesi. Ichthyological Exploration of Fresh- excelsus, Monacanthidae) to the at least 3-m waters 9, 311–23. long, 2000-kg ocean sunfish (Mola mola, Molidae). Ahlberg, P.E. and Johanson, Z. (1998) Osteolepiforms and Tetraodontiforms are distinctive in having a small the ancestry of tetrapods. Nature 395, 792–4. slit-like gill opening, and scales usually modified Albert, J.S. and Campos-da-Paz, R. (1998) Phylogenetic systematics of Gymnotiformes with diagnoses of 58 as spines, shields or plates. Additional characters clades: a review of available data. In: L.R. Malabarba, defining the order and relationships within it are R.E. Reis, R.P. Vari, Z.M.S. Lucena and C.A.S. Lucenca variously discussed by Winterbottom (1974), Tyler (eds) Phylogeny and Classification of Neotropical (1980), Leis (1984) and Tyler and Sorbini (1996). Fishes. Porto Alegre, Brazil: Edipucrs, pp. 419–60. Phylogeny and Systematics 37

Allen, G.R. (1980) A generic classification of the rain- fishes of the world (order Gadiformes). An annotated bowfishes (family Melanotaeniidae). Records of the and illustrated catalogue of cods, hakes, grenadiers and Western Australian Museum 8, 449–90. other gadiform fishes known to date. FAO Fisheries Arratia, G. (1998) Basal teleosts and teleostean phylo- Synopsis (125) 10, 1–442. geny: response to C. Patterson. Copeia 1998, 1109–13. Collette, B.B., McGowen, G.E., Parin, N.V. and Mito, S. Baldwin, C.C. and Johnson, G.D. (1996) Interrelation- (1984) Beloniformes: development and relationships. ships of Aulopiformes. In: M.L.J. Stiassny, L.R. Parenti In: H.G. Moser, W.J. Richards, D.M. Cohen, M.P. and G.D. Johnson (eds) Interrelationships of Fishes. Fahay, A.W. Kendall Jr and S.L. Richardson (eds) San Diego, CA: Academic Press, pp. 355–404. Ontogeny and Systematics of Fishes. American Bell, M.A. and Foster, S.A. (eds) (1994) The Evolutionary Society of Ichthyologists and Herpetologists Special Biology of the Threespine Stickleback. Oxford: Oxford Publication No. 1. Lawrence, KS: Allen Press, pp. University Press. 335–54. Bemis, W.E., Burggren, W. and Kemp, N.E. (eds) (1987) Compagno, L.J.V. (1988) Sharks of the Order Car- The biology and evolution of lungfishes: based on pro- charhiniformes. Princeton, NJ: Princeton University ceedings of a symposium held during the American Press. Society of Zoologists meeting in Denver, Colorado, Compagno, L.J.V. (1990) Relationships of the meg- December 27, 1984. Journal of Morphology Supple- amouth shark, Megachasma pelagios (Lamniformes; ment 1, 1–383. Megachasmidae), with comments on its feeding Bertelsen, E. (1984) Ceratioidei: development and rela- habits. In: H.L. Pratt Jr, S.H. Gruber and T. Taniuchi tionships. In: H.G. Moser, W.J. Richards, D.M. Cohen, (eds) Elasmobranchs as Living Resources: Advances in M.P. Fahay, A.W. Kendall Jr and S.L. Richardson (eds) Biology, Ecology, Systematics and the Status of the Ontogeny and Systematics of Fishes. American Fisheries. National Oceanic and Atmospheric Admin- Society of Ichthyologists and Herpetologists Special istration Technical Report National Marine Fisheries Publication No. 1. Lawrence, KS: Allen Press, pp. Service 90. Washington, DC: US Department of 325–34. Commerce, pp. 357–79. Birstein, V.J., Waldman, J.R. and Bemis, W.E. (eds) (1997) Cooper, J.A. and Chapleau, F. (1998) Phylogenetic status Sturgeon biodiversity and conservation. Environmen- of Paralichthodes algoensis (Pleuronectiformes: tal Biology of Fishes 48, 1–437. Parlichthodidae). Copeia 1998, 477–81. Britz, R. (1997) Egg surface structure and larval cement Costa, W.J.E.M. (1998) Phylogeny and classification of glands in nandid and badid fishes with remarks on the Cyprinodontiformes (Euteleostei: Atherinomor- phylogeny and zoogeography. American Museum pha): a reappraisal. In: L.R. Malabarba, R.E. Reis, R.P. Novitates No. 3195, 1–17. Vari, Z.M.S. Lucena and C.A.S. Lucenca (eds) Phylo- Buckup, P.A. (1998) Relationships of the Characidiinae geny and Classification of Neotropical Fishes. Porto and phylogeny of characiform fishes (Teleostei: Ostar- Alegre, Brazil: Edipucrs, pp. 534–60. iophysi). In: L.R. Malabarba, R.E. Reis, R.P. Vari, Z.M.S. de Carvalho, M.R. (1996) Higher-level elasmobranch Lucena and C.A.S. Lucenca (eds) Phylogeny and Clas- phylogeny, basal squaleans, and paraphyly. In: M.L.J. sification of Neotropical Fishes. Porto Alegre, Brazil: Stiassny, L.R. Parenti and G.D. Johnson (eds) Interrela- Edipucrs, pp. 123–44. tionships of Fishes. San Diego, CA: Academic Press, Carpenter, K.E., Collette, B.B. and Russo, J.L. (1995) pp. 35–62. Unstable and stable classifications of scombroid de Pinna, M.C.C. (1991) Concepts and tests of homology fishes. Bulletin of Marine Science 56, 379–405. in the cladistic paradigm. Cladistics 7, 367–94. Chapleau, F. (1993) Pleuronectiform relationships: a de Pinna, M.C.C. (1996) Teleostean monophyly. In: cladistic reassessment. Bulletin of Marine Science 52, M.L.J. Stiassny, L.R. Parenti and G.D. Johnson (eds) 516–640. Interrelationships of Fishes. San Diego, CA: Academic Cloutier, R. and Ahlberg, P.E. (1996) Morphology, charac- Press, pp. 147–62. ters, and the interrelationships of basal sarcoptery- de Pinna, M.C.C. (1998) Phylogenetic relationships of gians. In: M.L.J. Stiassny, L.R. Parenti and G.D. neotropical Siluriformes (Teleostei: Ostariophysi): Johnson (eds) Interrelationships of Fishes. San Diego, historical overview and synthesis of hypotheses. In: CA: Academic Press, pp. 445–79. L.R. Malabarba, R.E. Reis, R.P. Vari, Z.M.S. Lucena and Cohen, D.M. (ed.) (1989) Papers on the Systematics of C.A.S. Lucenca (eds) Phylogeny and Classification of Gadiform Fishes. Natural History Museum of Los Neotropical Fishes. Porto Alegre, Brazil: Edipucrs, pp. Angeles County Science Series, No. 32. 279–330. Cohen, D.M., Inada, T., Iwamoto, T. and Scialabba, N. de Queiroz, K. (1998) The general lineage concept of (1990) FAO species catalogue. Volume 10. Gadiform species, species criteria, and the process of speciation: a 38 Chapter 2

conceptual unification and terminological recommen- Stiassny, L.R. Parenti and G.D. Johnson (eds) Interrela- dations. In: D.J. Howard and S.H. Berlocher (eds) tionships of Fishes. San Diego, CA: Academic Press, Endless Forms: Species and Speciation. Oxford: pp. 117–46. Oxford University Press, pp. 57–75. Gill, A.C. (1996) Comments on an intercalar path for the Didier, D.A. (1995) Phylogenetic systematics of extant glossopharyngeal (cranial IX) nerve as a synapomorphy chimaeroid fishes (Holocephali, Chimaeroidei). of the Paracanthopterygii and on the phylogenetic po- American Museum Novitates No. 3119, 1–86. sition of the Gobiesocidae (Teleostei: Acanthomor- Dingerkus, G. (1986) Interrelationships of orectolobi- pha). Copeia 1996, 1022–9. form sharks (Chondrichthyes: Selachii). In: T. Uyeno, Gill, A.C. (1999) Subspecies, geographic forms and wide- R. Arai, T. Taniuchi and K. Matsuura (eds) Indo-Pacific spread Indo-Pacific coral-reef fish species: a call for Fish Biology. Proceedings of the Second International change in taxonomic practice. In: B. Seret and J.-Y. Sire Conference on Indo-Pacific Fishes. Tokyo: The Ichthy- (eds) Proceedings of the Fifth Indo-Pacific Fish Confer- ological Society of Japan, pp. 227–45. ence, Nouméa, 3–8 November 1997. Paris: Société Dingerkus, G. and Uhler, L.D. (1977) Enzyme clearing of Française d’Ichtyologie and Institut de Recherche pour alcian blue stained whole vertebrates for demonstra- le Développement, pp. 79–87. tion of cartilage. Stain Technology 52, 229–32. Gill, A.C. and Kemp, J.M. (in press) Widespread Indo- Dyer, B.S. and Chernoff, B. (1996) Phylogenetic relation- Pacific shore-fish species: a challenge for taxonomists, ships among atheriniform fishes (Teleostei: Atheri- biogeographers, ecologists, fishery and conservation nomorpha). Zoological Journal of the Linnean Society managers. Environmental Biology of Fishes. 117, 1–69. Gill, A.C. and Mooi, R.D. (1993) Monophyly of the Gram- Erdmann, M.V., Caldwell, R.L. and Moosa, M.K. (1998) matidae and of the Notograptidae, with evidence for Indonesian ‘King of the Sea’ discovered. Nature 395, their phylogenetic positions among perciforms. Bul- 335. letin of Marine Science 52, 327–50. Eschmeyer, W.N. (1990) Genera in a classification. In: Gill, T. (1890) The characteristics of the Dacty- W.N. Eschmeyer (ed.) Catalog of the Genera of Recent lopteroidea. Proceedings of the United States Na- Fishes. San Francisco, CA: California Academy of tional Museum 13, 243–8. Sciences, pp. 435–95. Grande, L. and Bemis, W.E. (1996) Interrelationships of Eschmeyer, W.N. (1997) A new species of Dactylopteri- Acipenseriformes, with comments on ‘Chondrostei’. dae (Pisces) from the Philippines and Australia, with a In: M.L.J. Stiassny, L.R. Parenti and G.D. Johnson (eds) brief synopsis of the family. Bulletin of Marine Science Interrelationships of Fishes. San Diego, CA: Academic 60, 727–38. Press, pp. 85–115. Ferraris, C.J. Jr and de Pinna, M.C.C. (1999) Higher-level Grande, L. and Bemis, W.E. (1998) A comprehensive names for catfishes (Actinopterygii: Ostariophysi: phylogenetic study of amiid fishes (Amiidae) based on Siluriformes). Proceedings of the California Academy comparative skeletal anatomy. An empirical search for of Sciences 51, 1–17. interconnected patterns of natural history. Journal of Fink, S.V. and Fink, W.L. (1981) Interrelationships of the Vertebrate Paleontology Supplement 8, 1–690. ostariophysan fishes (Teleostei). Zoological Journal of Greenwood, P.H. and Rosen, D.E. (1971) Notes on the the Linnean Society 78, 199–286. structure and relationships of the alepocephaloid Fink, S.V. and Fink, W.L. (1996) Interrelationships of fishes. American Museum Novitates No. 2473, 1–41. ostariophysan fishes (Teleostei). In: M.L.J. Stiassny, Greenwood, P.H., Rosen, D.E., Weitzman, S.H. and L.R. Parenti and G.D. Johnson (eds) Interrelationships Myers, G.S. (1966) Phyletic studies of teleostean of Fishes. San Diego, CA: Academic Press, pp. 209–49. fishes, with a provisional classification of living Forey, P.L. (1998) History of the Coelacanth Fishes. forms. Bulletin of the American Museum of Natural London: Chapman & Hall. History 131, 339–456. Forey, P.L., Littlewood, D.T.J., Ritchie, P. and Meyer, A. Greenwood, P.H., Miles, R.S. and Patterson, C. (eds) (1996) Interrelationships of elopomorph fishes. In: (1973) Interrelationships of Fishes. London: Academic M.L.J. Stiassny, L.R. Parenti and G.D. Johnson (eds) Press. Interrelationships of Fishes. San Diego, CA: Academic Harold, A.S. and Weitzman, S.H. (1996) Interrelation- Press, pp. 175–91. ships of Stomiiform fishes. In: M.L.J. Stiassny, L.R. Gardiner, B.G. and Schaeffer, B. (1989) Interrelationships Parenti and G.D. Johnson (eds) Interrelationships of of lower actinopterygian fishes. Zoological Journal of Fishes. San Diego, CA: Academic Press, pp. 333–53. the Linnean Society 97, 135–87. Helfman, G.S., Collette, B.B. and Facey, D.E. (1997) The Gardiner, B.G., Maisey, J.G. and Littlewood, D.T.J. (1996) Diversity of Fishes. Malden, MA: Blackwell Science. Interrelationships of basal neopterygians. In: M.L.J. Hennig, W. (1950) Grundzüge einer Theorie der Phylogeny and Systematics 39

phylogenetischen Systematik. Berlin: Deutscher other look. In: American Society of Ichthyologists Zentralverlag. and Herpetologists, 77th Annual Meeting, Program Hennig, W. (1966) Phylogenetic Systematics. Urbana: and Abstracts for 1997. Seattle, WA: University of University of Illinois Press. Washington, p. 176. Howes, G.J. (1991) Biogeography of gadoid fishes. Journal Kotlyar, A.N. (1996) Beryciform Fishes of the World of Biogeography 18, 595–622. Ocean. Moscow: VNIRO Publishing. [In Russian] Howes, G.J. (1992) Notes on the anatomy and classifica- Kottelat, M. (1997) European freshwater fishes. An tion of ophidiiform fishes with particular reference to heuristic checklist of the freshwater fishes of Europe the abyssal genus Acanthonus Günther, 1878. Bulletin (exclusive of former USSR), with an introduction for of the Natural History Museum, London, Zoology non-systematists and comments on nomenclature and Series, 58, 95–131. conservation. Biologia 52, 1–271. Imamura, I. (2000) An alternative hypothesis on the Kurlansky, M. (1997) Cod: a Biography of the Fish That phylogenetic position of the family Dactylopteridae Changed the World. New York: Penguin Books. (Pisces: Teleostei), with a proposed new classification. Last, P.R. and Stevens, J.D. (1994) Sharks and Rays of Ichthyological Research 47, 203–22. Australia. Australia: CSIRO. Imamura, I. and Shinohara, G. (1998) Scorpaeniform fish Lauder, G.V. and Liem, K.F. (1983) The evolution and phylogeny: an overview. Bulletin of the National Sci- interrelationships of the actinopterygian fishes. Bul- ence Museum, Tokyo, Series A 24, 185–212. letin of the Museum of Comparative Zoology 150, Jamieson, B.G.M. (1991) Fish Evolution and Syste- 95–197. matics: Evidence From Spermatozoa. Cambridge: Lecointre, G. and Nelson, G. (1996) Clupeomorpha, Cambridge University Press. sister group of Ostariophysi. In: M.L.J. Stiassny, L.R. Janvier, P. (1981) The phylogeny of the Craniata, with Parenti and G.D. Johnson (eds) Interrelationships of particular reference to the significance of fossil Fishes. San Diego, CA: Academic Press, pp. 193–207. ‘agnathans’. Journal of Vertebrate Paleontology 1, Leis, J.M. (1984) Tetraodontiformes: relationships. In: 121–59. H.G. Moser, W.J. Richards, D.M. Cohen, M.P. Fahay, Janvier, P. (1996) The dawn of vertebrates: characters ver- A.W. Kendall Jr and S.L. Richardson (eds) Ontogeny sus common ascent in the rise of the current vertebrate and Systematics of Fishes. American Society of phylogenies. Paleontology 39, 250–87. Ichthyologists and Herpetologists Special Publication Johnson, G.D. (1992) Monophyly of the euteleostean No. 1. Lawrence, KS: Allen Press, pp. 459–63. clades Neoteleostei, Eurypterygii and Ctenosqua- Li, G.-Q. and Wilson, M.V.H. (1996) Phylogeny of mata. Copeia 1992, 8–25. Osteoglossomorpha. In: M.L.J. Stiassny, L.R. Parenti Johnson, G.D. (1993) Percomorph phylogeny: progress and G.D. Johnson (eds) Interrelationships of Fishes. and problems. Bulletin of Marine Science 52, 3–28. San Diego, CA: Academic Press, pp. 163–74. Johnson, G.D. and Anderson, W.D. Jr (eds) (1993) Pro- Liem, K.F. (1998) Swampeels and their allies. In: J.R. ceedings of the Symposium on Phylogeny of the Perco- Paxton and W.N. Eschmeyer (eds) Encyclopedia of morpha, June 15–17, 1990, held in Charleston, South Fishes, 2nd edn. San Diego, CA: Academic Press, pp. Carolina at the 70th Annual Meetings of the American 173–4. Society of Ichthyologists and Herpetologists. Bulletin Lovejoy, N.R. (2000) Reinterpreting recapitulation: sys- of Marine Science 52, 1–626. tematics of needlefishes and their allies (Teleostei: Johnson, G.D. and Gill, A.C. (1998) Perches and their Beloniformes). Evolution 54, 1349–62. allies. In: J.R. Paxton and W.N. Eschmeyer (eds) Ency- Lundberg, J.G., Kottelat, M., Smith, G.R., Stiassny, clopedia of Fishes, 2nd edn. San Diego, CA: Academic M.L.J. and Gill, A.C. (2000) So many fishes, so little Press, pp. 181–94. time: an overview of recent ichthyological discovery in Johnson, G.D. and Patterson, C. (1993) Percomorph phy- continental waters. Annals of the Missouri Botanical logeny: a survey of acanthomorphs and a new proposal. Garden 87, 26–62. Bulletin of Marine Science 52, 554–626. Maisey, J.G. (1985) Relationships of the megamouth Johnson, G.D. and Patterson, C. (1996) Relationships of shark, Megachasma. Copeia 1985, 228–31. lower euteleostean fishes. In: M.L.J. Stiassny, L.R. Malabarba, L.R., Reis, R.E., Vari, R.P., Lucena, Z.M.S. and Parenti and G.D. Johnson (eds) Interrelationships of Lucenca, C.A.S. (eds) (1998) Phylogeny and Classifi- Fishes. San Diego, CA: Academic Press, pp. 251–332. cation of Neotropical Fishes. Porto Alegre, Brazil: Johnson, G.D. and Patterson, C. (1997) The gill arches of Edipucrs. gonorynchiform fishes. South African Journal of Sci- Markle, D.F. (1989) Aspects of character homology and ence 93, 594–600. phylogeny of the Gadiformes. In: D.M. Cohen (ed.) Johnson, G.D. and Springer, V.G. (1997) Elassoma: an- Papers on the Systematics of Gadiform Fishes. 40 Chapter 2

Natural History Museum of Los Angeles County Sci- Nelson, G.J. (1989a) Species and taxa: systematics and ence Series, No. 32, pp. 59–88. evolution. In: D. Otte and J.A. Endler (eds) Speciation Mayden, R.L. (1997) A hierarchy of species concepts: the and its Consequences. Sunderland, MA: Sinauer, pp. denouement in the saga of the species problem. In: M.F. 60–81. Claridge, H.A. Dawah and M.R. Wilson (eds) Species: Nelson, G.J. (1989b) Phylogeny of major fish groups. The Units of Biodiversity. London: Chapman & Hall, In: B. Fernholm, K. Bremer and H. Jörnvall (eds) The pp. 381–424. Hierarchy of Life. Molecules and Morphology in McEachran, J.D., Dunn, K.A. and Miyake, T. (1996) Inter- Phylogenetic Analysis. Amsterdam: Elsevier Science relationships of the batoid fishes (Chondrichthyes: Publishers, pp. 325–36. Batoidei). In: M.L.J. Stiassny, L.R. Parenti and G.D. Nelson, G.J. (1994) Homology and systematics. In: B.K. Johnson (eds) Interrelationships of Fishes. San Diego, Hall (ed.) Homology: The Hierarchical Basis of Com- CA: Academic Press, pp. 63–84. parative Biology. San Diego, CA: Academic Press, pp. McLennan, D.A., Brooks, D.R. and McPhail, J.D. (1988) 101–49. The benefits of communication between comparative Nelson, G.J. and Platnick, N.I. (1981) Systematics and ethology and phylogenetic systematics: a case study Biogeography: Cladistics and Vicariance. New York: using gasterosteid fishes. Canadian Journal of Zoology Columbia University Press. 66, 2177–90. Nelson, J.S. (1976) Fishes of the World. New York: John Miya, M. and Nishida, M. (1996) Molecular phylogenetic Wiley & Sons. perspective of the evolution of the deep-sea fish genus Nelson, J.S. (1984) Fishes of the World, 2nd edn. New Cyclothone (Stomiiformes: Gonostomatidae). Ichthy- York: John Wiley & Sons. ological Research 43, 375–98. Nelson, J.S. (1994) Fishes of the World, 3rd edn. New Mooi, R.D. (1990) Egg surface morphology of pseudo- York: John Wiley & Sons. chromoids (Perciformes: Percoidei), with comments Nelson, J.S. and Hart, P.J.B. (eds) (1999) The species con- on phylogenetic implications. Copeia 1990, 455–75. cept in fish biology. Reviews in Fish Biology and Fish- Mooi, R.D. and Gill, A.C. (1995) Association of epaxial eries 9, 277–382. musculature with dorsal-fin pterygiophores in acan- Nielsen, J.G., Cohen, D.M., Markle, D.F. and Robins, thomorph fishes, and its phylogenetic significance. C.R. (1999) FAO species catalogue. Volume 18. Ophidi- Bulletin of the Natural History Museum, London, iform fishes of the world (order Ophidiiformes). An Zoology Series 61, 121–37. annotated and illustrated catalogue of pearlfishes, Mooi, R.D. and Johnson, G.D. (1997) Dismantling the cusk-eels, brotulas and other ophidiiform fishes Trachinoidei: evidence of a scorpaenoid relationship known to date. FAO Fisheries Synopsis (125) 18, 1–178. for the Champsodontidae. Ichthyological Research 44, Norman, J.R. and Fraser, F.C. (1937) Giant Fishes, 143–76. Whales and Dolphins. London: Putman. Moore, J.A. (1993) Phylogeny of the Trachichthyiformes Olney, J.E., Johnson, G.D. and Baldwin, C.C. (1993) Phy- (Teleostei: Percomorpha). Bulletin of Marine Science logeny of the lampridiform fishes. Bulletin of Marine 52, 114–36. Science 52, 137–69. Moser, H.G., Richards, W.J., Cohen, D.M., Fahay, M.P., Parenti, L.R. (1981) A phylogenetic and biogeographic Kendall, A.W. Jr and Richardson, S.L. (eds) (1984) analysis of cyprinodontiform fishes (Teleostei, Atheri- Ontogeny and Systematics of Fishes. American nomorpha). Bulletin of the American Museum of Society of Ichthyologists and Herpetologists Special Natural History 168, 335–557. Publication No. 1. Lawrence, KS: Allen Press. Parenti, L.R. (1984) On the relationships of phallostethid Murray, A.M. and Wilson, M.V.H. (1999) Contributions fishes (Atherinomorpha), with notes on the anatomy of fossils to the phylogenetic relationships of the per- of Phallostethus dunckeri Regan, 1913. American copsiform fishes (Teleostei: Paracanthopterygii): order Museum Novitates No. 2779, 1–12. restored. In: G. Arratia and H.P. Schultz (eds) Mesozoic Parenti, L.R. (1986) The phylogenetic significance of Fishes 2: Systematics and Fossil Record. München: bone types in euteleost fishes. Zoological Journal of Verlag Dr Friedrich Pfeil, pp. 397–411. the Linnean Society 87, 37–51. Nelson, G.J. (1969) Gill arches and the phylogeny of Parenti, L.R. (1993) Relationships of atherinomorph fishes, with notes on the classification of vertebrates. fishes (Teleostei). Bulletin of Marine Science 52, Bulletin of the American Museum of Natural History 170–96. 141, 475–552. Parenti, L.R. and Song, J. (1996) Phylogenetic signifi- Nelson, G.J. (1972) Comments on Hennig’s ‘Phyloge- cance of the pectoral–pelvic association in acantho- netic Systematics’ and its influence on ichthyology. morph fishes: a reassessment using comparative Systematic Zoology 21, 364–74. neuroanatomy. In: M.L.J. Stiassny, L.R. Parenti and Phylogeny and Systematics 41

G.D. Johnson (eds) Interrelationships of Fishes. San Asia. Proceedings of the California Academy of Sci- Diego, CA: Academic Press, pp. 427–44. ences 42, 295–302. Patterson, C. (1982) Morphological characters and Robins, C.R. (1997) Review of Interrelationships of homology. In: K.A. Joysey and A.E. Friday (eds) Prob- fishes, edited by M.L.J. Stiassny, L.R. Parenti and G.D. lems of Phylogenetic Reconstruction. Systematic As- Johnson. Bulletin of Marine Science 61, 937–9. sociation Special Volume No. 25. London: Academic Rosen, D.E. (1984) Zeiformes as primitive plectognath Press, pp. 21–74. fishes. American Museum Novitates, No. 2782, 1–45. Patterson, C. (1988) Homology in classical and molecular Rosen, D.E. (1985) An essay on euteleostean classifica- biology. Molecular Biology and Evolution 5, 603–25. tion. American Museum Novitates, No. 2827, 1–57. Patterson, C. (1998) Comments on basal teleosts and Rosen, D.E. and Parenti, L.R. (1981) Relationships of teleostean phylogeny, by Gloria Arratia. Copeia 1998, Oryzias, and the groups of atherinomorph fishes. 1107–9. American Museum Novitates No. 2719, 1–25. Patterson, C. and Rosen, D.E. (1977) Review of ichthyo- Rosen, D.E. and Patterson, C. (1969) The structure and dectiform and other Mesozoic teleost fishes and the relationships of the paracanthopterygian fishes. Bul- theory and practice of classifying fossils. Bulletin letin of the American Museum of Natural History 141, of the American Museum of Natural History 158, 357–474. 81–172. Saeed, B., Ivantsoff, W. and Crowley, L. (1994) Systematic Patterson, C. and Rosen, D.E. (1989) The Paracan- relationships of the atheriniform families within Divi- thopterygii revisited: order and disorder. In: D.M. sion I of the Series Atherniomorpha (Acanthopterygii) Cohen (ed.) Papers on the Systematics of Gadiform with relevant historical perspectives. Voprosy Ikhti- Fishes. Natural History Museum of Los Angeles ologii 34, 1–32. County Science Series, No. 32, pp. 5–36. Secor, D.H., Arefjev, V., Nikolaev, A. and Sharov, Paxton, J.R. and Eschmeyer, W.N. (eds) (1994) Encyclope- A. (2000) Restoration of sturgeons: lessons from the dia of Fishes. Sydney: University of New South Wales Caspian Sea Sturgeon Ranching Programme. Fish and Press. Fisheries 1, 215–30. Paxton, J.R. and Eschmeyer, W.N. (eds) (1998) Encyclo- Shirai, S. (1992) Squalean Phylogeny: a New Framework pedia of Fishes, 2nd edn. San Diego, CA: Academic of ‘Squaloid’ Sharks and Related Taxa. Sapporo, Japan: Press. Hokkaido University Press. Pietsch, T.W. (1984) Lophiiformes: development and Shirai, S. (1996) Phylogenetic interrelationships of relationships. In: H.G. Moser, W.J. Richards, D.M. neoselachians (Chondrichthyes: Euselachii). In: M.L.J. Cohen, M.P. Fahay, A.W. Kendall Jr and S.L. Stiassny, L.R. Parenti and G.D. Johnson (eds) Interrela- Richardson (eds) Ontogeny and Systematics of Fishes. tionships of Fishes. San Diego, CA: Academic Press, American Society of Ichthyologists and Herpetologists pp. 9–34. Special Publication No. 1. Lawrence, KS: Allen Press, Siebert, D.J. (1997) Notes on the anatomy and relation- pp. 320–5. ships of Sundasalanx (Teleostei: Clupeidae), with the Pietsch, T.W. and Anderson, W.D. Jr (eds) (1997) description of new species. Bulletin of the Natural Collection Building in Ichthyology and Herpetology. History Museum, London, Zoology Series 63, 13– American Society of Ichthyologists and Herpetologists 26. Special Publication No. 3. Lawrence, KS: Allen Springer, V.G. and Fraser, T.H. (1976) Synonymy of the Press. fish families Cheilobranchidae (= Alabetidae) and Pietsch, T.W. and Grobecker, D.B. (1987) Frogfishes of the Gobiesocidae, with descriptions of two new species of World: Systematics, Zoogeography, and Behavioral Alabes. Smithsonian Contributions to Zoology 216, Ecology. Stanford, CA: Stanford University Press. 1–43. Randall, J.E. (1998) Zoogeography of shore fishes of the Stiassny, M.L.J. (1986) The limits and relationships of the Indo-Pacific region. Zoological Studies 37, 227–68. acanthomorph teleosts. Journal of Zoology (London) Rasmussen, A.-S. and Arnason, U. (1999) Phylogenetic (B) 1, 411–60. studies of complete mitochrondrial DNA molecules Stiassny, M.L.J. (1990) Notes on the anatomy and rela- place cartilaginous fishes within the tree of bony tionships of the bedotiid fishes of Madagascar, with a fishes. Journal of Molecular Evolution 48, 118–23. taxonomic revision of the genus Rheocles (Atheri- Roberts, C.D. (1993) Comparative morphology of spined nomorpha: Bedotiidae). American Museum Novitates scales and their phylogenetic significance in the No. 2993, 1–14. Teleostei. Bulletin of Marine Science 52, 60–113. Stiassny, M.L.J. (1996) Basal ctenosquamate relation- Roberts, T.R. (1981) Sundasalangidae, a new family of ships and the interrelationships of the Myctophiform minute freshwater salmoniform fishes from Southeast (Scopelomorph) fishes. In: M.L.J. Stiassny, L.R. Parenti 42 Chapter 2

and C.D. Johnson (eds) Interrelationships of Fishes. Vari, R.P. (1998) Higher level phylogenetic concepts San Diego, CA: Academic Press, pp. 405–26. within characiforms (Ostariophysi): a historical re- Stiassny, M.L.J., Parenti, L.R. and Johnson, G.D. (eds) view. In: L.R. Malabarba, R.E. Reis, R.P. Vari, Z.M.S. (1996) Interrelationships of Fishes. San Diego, CA: Lucena and C.A.S. Lucenca (eds) Phylogeny and Clas- Academic Press. sification of Neotropical Fishes. Porto Alegre, Brazil: Taylor, W.R. (1967) An enzyme method of clearing and Edipucrs, pp. 111–22. staining small vertebrates. Proceedings of the United White, B.N., Lavenberg, R.J. and McGowen, G.E. (1984) States National Museum 122(3596): 1–17. Atheriniformes: development and relationships. In: Taylor, W.R. and Van Dyke, G.C. (1985) Revised proce- H.G. Moser, W.J. Richards, D.M. Cohen, M.P. Fahay, dures for staining and clearing small fishes and other A.W. Kendall Jr and S.L. Richardson (eds) Ontogeny vertebrates for bone and cartilage study. Cybium 9, and Systematics of Fishes. American Society of 107–19. Ichthyologists and Herpetologists Special Publication Tyler, J.C. (1980) Osteology, phylogeny, and higher No. 1. Lawrence, KS: Allen Press, pp. 355–62. classification of the fishes of the order Plectognathi Wiley, E.O. (1976) The phylogeny and biogeography of (Tetraodontiformes). National Oceanic and Atmos- fossil and Recent gars (Actinopterygii: Lepisosteidae). pheric Administration Technical Report, National University of Kansas, Museum of Natural History, Marine Fisheries Service Circular No. 434, 1–422. Miscellaneous Publication 64, 1–111. Tyler, J.C. and Sorbini, L. (1996) New superfamily and Winterbottom, R. (1974) The familial phylogeny of the three new families of tetraodontiform fishes from the Tetraodontiformes (Acanthopterygii: Pisces) as evi- Upper Cretaceous: the earliest and most morphologi- denced by their comparative myology. Smithsonian cally primitive plectognaths. Smithsonian Contribu- Contributions to Zoology 155, 1–201. tions to Paleobiology 82, 1–59. 3 Historical Biogeography of Fishes

R.D. MOOI AND A.C. GILL

3.1 INTRODUCTION two approaches. Brown and Lomolino (1998) pro- vide an excellent university-level text for all as- Biogeography is the study of the distribution of life pects of the field (historical and ecological). The on Earth, or which organisms live where and why. chapter by Jones et al. (Chapter 16, this volume) It is fundamental to understanding the evolution deals with ecological biogeography, but here we of species, the development of diversity, and the focus on the historical aspect of biogeography as a composition and interaction of organisms within companion to the preceding chapter on the phy- the environment. At its most basic, biogeography logeny of fishes. We briefly outline the concepts provides distributional and faunal composition and methods before summarizing fish distribu- data that directly impact fishery biologists and tions, faunal composition, and historical biogeog- conservationists, where understanding of distri- raphy of freshwater and marine regions. bution and species limits will determine resource use. In this chapter we convey a more complete 3.1.1 What is historical view of biogeography as a dynamic and exciting biogeography? conceptual science that generates hypotheses and makes testable predictions about the natural Historical biogeographers rely on accurate distri- world. butional information, good taxonomy and well- Biogeography is divided into two components. corroborated phylogenies to identify distribution Historical biogeography reconstructs the histori- patterns and form robust hypotheses. Comprehen- cal influences on the distribution of organisms sive knowledge of distributions of most organ- over long temporal and often large spatial scales, isms, including fishes, is not generally available. thereby identifying patterns of distribution and Hence, biogeographic hypotheses are often built processes that might explain them. Ecological bio- on incomplete data. Our understanding of fish geography accounts for distribution in terms of diversity and distribution has changed dramati- the interaction of organisms with their physical cally over the last three decades with increased col- and biotic environment over short temporal and lecting effort and improved methods. The number small spatial scales, thereby exploring processes of described fish species increases by about 200 per that limit distribution and examining patterns of year (Eschmeyer 1998), due to more thorough species richness. Myers and Giller (1988) proposed exploration of poorly known areas or habitats, that a complete understanding of distributions of and recognition of what were formerly widespread organisms cannot be attained without knowledge ‘species’ as pairs or complexes of allopatric species of the full spectrum of ecological and historical with separate evolutionary histories (Gill 1999). In processes. Their book is an attempt to integrate the either case, distributional ranges are altered and 44 Chapter 3 patterns change. This emphasizes the importance mary freshwater fishes, being confined to particu- of continued broad faunal surveys, museum col- lar watersheds, provided the most important data lections and monographic work. Effective bio- for biogeography continues to influence taxon geographic studies are based on a synthesis of the choice in fish biogeographic studies. distributions of several groups or repeating dis- The modern era of biogeography, beginning tributional patterns within a taxon; historical bio- about 1960, has as a fundamental principle that geography relies on common patterns. Earth and life have evolved together. This concept was bolstered by the simultaneous re-emergence among geologists of support for continental drift 3.2 CONCEPTS AND and plate tectonic theory. This new, unstable Earth METHODS required that the foundations of historical biogeog- 3.2.1 Past influences on raphy be examined closely. Through a reinterpre- today’s approaches tation of Croizat’s (1964) work in a phylogenetic framework, Nelson, Rosen (ichthyologists) and The history of biogeography is fascinating, and is Platnick (arachnologist) revolutionized biogeo- required background to comprehend present-day graphy. They developed a method where cladistic polemics. Detailed accounts can be found in summaries of organismal relationships and their Nelson (1978) and Humphries and Parenti (1986, distributions could be compared for congruence 1999) (cladistic perspective), and Mayr (1982, with geological area relationships as formulated pp. 439–55) (classical perspective). Brown and by geologists; biotic data could provide a direct Lomolino (1998) provide a balanced account of means to test hypotheses of geohistory (Platnick the development of ecological and historical bio- and Nelson 1978; Nelson and Platnick 1981). geography. A summary of the last 20 years can be Ichthyologists played a leading role in the biogeog- found in Holloway and Hall (1998). raphy revolution (e.g. G. Nelson, Rosen, Patterson, Early biogeography was dominated by the iden- Parenti, Wiley). tification of faunal realms, some of which are still recognized. Sclater (1858) divided the Earth into 3.2.2 Pattern and process: six regions based on the distribution of birds (Fig. congruence and the roles of 3.1a), later applied to animals in general by Wallace vicariance and dispersal (1876), and Forbes (1856) produced a map dividing the oceans into provinces. Until the late 20th cen- Historical biogeography is essentially a search for tury, the geography of the Earth was considered an explanation of some form of endemic distribu- stable over the evolution of distributions. Hence, tion: why is a taxon restricted to a particular biogeographers had no mechanism other than dis- geographic area? The field’s development can be persal to explain them. Higher taxa were hypothe- traced through the philosophical dominance of sized to arise in a centre of origin from which one of two main processes to explain endemism: new species disperse as widely as their abilities vicariance and dispersal. In a vicariance explana- and suitable environment allow, displacing more tion, taxa are endemic because their ancestors primitive species towards peripheral areas. A originally occurred in an area and their descen- comprehensive dispersalist summary of fresh- dants remain. An ancestral population is divided water fish distribution was provided by Darlington when a barrier appears that individuals cannot (1957). negotiate (Fig. 3.2a). Eventually, the two groups Fish biogeography became dominated by differentiate into taxa that are then endemic to Myers’ (1938) concept of primary, secondary and either side of the barrier. Here, the barrier is a cause tertiary freshwater fishes, a scheme that identified of the disjunction and the barrier and the taxa are of the degree of salt tolerance of fish taxa, with pri- equal age. In a dispersal explanation, endemic taxa mary being intolerant. The implication that pri- arise as the result of the dispersal into new areas by Historical Biogeography 45

(a)

Palaearctic Nearctic

Ethiopian Oriental

Neotropical

Australian

(b) Arctic Ocean

Eu EA NA Eu

In Pacific Atlantic Bo Bo CA Tethys Su Ocean Ocean Su NG Af Indian Ocean NG Af NC SAf Mad Aus/Tas NZ SA SAf

Southern Ocean

Fig. 3.1 Biogeographic realms. (a) World map showing the limits of Wallace’s six terrestrial biogeographic regions, still in use today. (b) A schematic of biogeographic regions based not on landmasses but ocean basins; each continent is inferred to be composite because it is part of the biogeographic regions of the oceans it contacts. The Tethys Sea is a body of water that separated the portions of Borneo, Sulawesi and New Guinea that later came in contact. This Tethys should not be confused with the larger Tethys hypothesized to have separated India from Asia. Af, Africa; Aus, Australia; Bo, Borneo; CA, Central America; EA, East Asia; Eu, Europe; In, India; Mad, Madagascar; NA, North America; NC, New Caledonia; NG, New Guinea; NZ, New Zealand; SA, South America; SAf, South Africa; Su, Sulawesi; Tas, Tasmania. (Source: a, from Nelson and Platnick 1984; reprinted by permission of Carolina Biological Supply Company. b, from Parenti 1991.) ancestors that originally occurred elsewhere (Fig. the roles of dispersal and vicariance in organismal 3.2b). In this case, the range of the ancestor is distribution. Early comments of Platnick and limited by a previously existing barrier that is later Nelson (1978, p. 8) should have made the process crossed by chance dispersal by some members of debate unnecessary: the population. Over time, surviving colonizers in the new area would evolve into a distinct taxon. the first question we should ask when con- Here, the barrier is older than the endemic taxa. fronted with the allopatric distribution of The central theme of modern historical bio- some group is not ‘Is this pattern the result geography, discovering congruent patterns of of vicariance or dispersal?’ but ‘Does this species and area relationships, has often been over- pattern correspond to a general pattern of shadowed by a focus on the relative importance of area interconnections (and thus reflect the 46 Chapter 3

Vicariance Dispersal (a) (b)

Ancestral population Ancestral population and barrier

Barrier appears Dispersal across barrier Fig. 3.2 Schematic of a vicariance explanation (a) and a dispersal explanation (b) of the disjunct distribution of two taxa (A1, A2). A1 A2 A1 A2 (Source: modified from Nelson and Platnick 1984; reprinted by permission of Carolina Biological Subsequent speciation Subsequent speciation Supply Company.)

history of those areas) or not? . . . After a hy- simony analysis of endemicity, introduced by pothesized general pattern is corroborated, Rosen (1984) and critiqued by Humphries (1989, we may be able to ascribe it to vicariance or 2000); and panbiogeography, reviewed by Craw dispersal by the use of independent evidence et al. (1999), with many critics (e.g. Seberg 1986; of earth history. Platnick and Nelson 1989; Mayden 1992a). For more detailed discussion of all methods, see Myers The development of methods to detect patterns of and Giller (1988), Morrone and Crisci (1995) and distribution and underlying history has been the Humphries and Parenti (1999). Humphries (2000) focus of theoretical biogeographers for the last 30 provides a historical perspective on biogeographic years. methods with a useful reference list. Two common methods used by ichthyologists are introduced 3.2.3 Methods below. There is no consensus on a preferred method of Phylogeography analysis; some workers have described historical biogeography as a subject looking for a method Phylogeography is a new subdiscipline that exam- (Tassy and Deleporte 1999). Most workers agree in ines the distribution of gene lineages within principle that such a method should include and among closely related species (Avise 2000). hypothesis testing based on a phylogenetic sys- Usually, genetic divergence is used to estimate tematic framework using area cladograms. Space relationships among populations and molecular does not permit even a brief overview of the main clocks to date lineages, with resulting unrooted methods in historical biogeography. Three that are similarity networks, called phenograms, used to not discussed further are dispersalism, largely dis- hypothesize species relationships, directions of credited but some principles such as ‘centres of ori- dispersal, number of invasions and vicariance (e.g. gin’ continue to be invoked implicitly, and even Bowen and Grant 1997; Bowen et al. 2001). explicitly (e.g. Bremer 1992; Briggs 1999a,b); par- The recent surge in interest in biogeography is Historical Biogeography 47 an important contribution of phylogeography, and ogical or climatic episodes. However, Page (1989) as methods are refined it will undoubtedly advance and Page and Lydeard (1994) have argued that com- understanding of fish distribution. Others will dis- mon patterns can be used to explore dispersal. agree but we recognize some fundamental issues There are two basic steps in cladistic biogeogra- which have yet to be fully resolved or addressed by phy, the construction of area cladograms from practitioners: taxon cladograms and the integration of these into 1 distributions are often examined on a case-by- a general area cladogram (Fig. 3.3). Area clado- case basis, emphasizing process over pattern, often grams are constructed by replacing terminal taxa imposing dispersal/centre of origin explanations; on an organismal cladogram with their distribu- 2 use of molecular clocks remains controversial, tions. This is simple if each taxon has a unique dis- since it has yet to be established that rates are tribution, but realistic situations with widespread constant; taxa (those found in more than one area) and redun- 3 frequent reliance on similarity matrices for phy- dant distributions (where two taxa occur in the logeny reconstruction is problematic (Farris 1981; same area) become complicated. Comparing area Murphy and Doyle 1998); cladograms is equally problematic. Attempts to 4 the dichotomous branching forced on the data address these problems have filled the pages of by tree-building algorithms does not reflect the several journals without consensus. reticulate nature of gene exchange at the popula- General area cladograms can be constructed tion level, hence the trees will not necessarily several ways, the common methods being compo- portray history of the populations or of their nent analysis, three-area statements, reconciled distributions. trees and Brook’s parsimony analysis (for details Despite these caveats, phylogeography has and see Myers and Giller 1988; Brooks and McClennan will contribute important data and hypotheses for 1991; Morrone and Crisci 1995; Humphries and historical biogeography. This will be particularly Parenti 1999). A fifth promising method is analysis true if it continues to move in the direction out- of paralogy-free subtrees (Nelson and Ladiges lined by Arbogast and Kenagy (2001) in using 1996). more comparative approaches, and more closely complementing or even incorporating cladistic biogeographic methods. Phylogeographic studies 3.3 DISTRIBUTION, FAUNAL of fishes are summarized in Avise (2000, pp. COMPOSITION AND 180–96). HISTORICAL BIOGEOGRAPHY Cladistic biogeography BY REGION Cladistic biogeography was introduced by Platnick and Nelson (1978), with complete exposi- The number of valid, named species of living fishes tions of the method in Nelson and Platnick (1981) is about 25000, accounting for just over half of ver- and Nelson and Rosen (1981) and a newer review in tebrate diversity. The actual number of species is Humphries and Parenti (1986, 1999). Cladistic bio- expected to approach 33000 or more with thor- geography uses organismal history reconstructed ough investigation of relatively unsurveyed local- using cladistics (see Gill and Mooi, Chapter 2, this ities and habitats (Vari and Malabarba 1998) and volume) to imply area relationships which reflect reappraisal of widespread species. Hence, specific geological and ecological history. This method numbers reported here will quickly become dated, uses implied area relationships to search for although the general character of the remarks will general patterns among taxa. General patterns remain reliable. speak to common processes, which are usually Somewhat less than 60% of fishes can be con- sought among vicariant events initiated by geol- sidered marine and almost 40% are principally 48 Chapter 3

(a)

A1 A2 A3

A1 A2 A3 (b) Dispersal hypothesis

AU AU Area cladogram SA or SA Taxon A1 A2 A3 Region SA AU NZ NZ NZ

Vicariance hypothesis

AU AU SA+AU+NZ SA + SA NZ NZ

Taxon cladograms Area cladograms (c) A1 A2 A3 SA AU NZ Congruence among General area cladograms cladogram B1 B2 B3 SA AU NZ SA AU NZ SA AU NZ

C1 C2 C3 SA AU NZ

Fig. 3.3 Historical explanations in biogeography. (a) Geographical distribution and cladogram of three species, A1, A2 and A3. (b) The cladogram can be translated into an area cladogram by replacing the taxa with their distributions. Corresponding dispersal and vicariance hypotheses are shown to account for the distribution. Note that even in the simplest instance (no ad hoc extinction or multiple colonization events), there are two dispersal hypotheses that account equally well for the known phylogenetic and area relationships but only one vicariance hypothesis. Even so, neither the dispersal nor vicariance hypothesis can be favoured; a unique example can have a unique distribution and history. (c) A cladistic biogeographic analysis with the construction of area cladograms for several taxonomic groups and derivation of a general area cladogram. A common pattern of history and distribution has been discovered that suggests a general explanation; the vicariance hypothesis of (b) best provides this explanation. (Source: a, modified from Morrone and Crisci 1995, reprinted by permission; b, modified from Brown and Lomolino 1998; c, modified from Morrone and Crisci 1995, reprinted by permission.) Historical Biogeography 49 freshwater residents. Less than 1% (about 250 Table 3.1 Estimates of numbers of fish species species) normally migrate between fresh and salt for freshwater and marine biogeographic regions. water (see Metcalfe et al., Chapter 8, this volume). Freshwater regional names follow Fig. 3.1. Marine estimates are for shorefishes only (defined as fishes Many people have remarked on the incredible bias occurring from shore to 200m depth). Almost all counts of fish diversity in favour of fresh water, with about are underestimates. Note that the historical reality for 10000 of 25000 species found in only 0.01% of the many of these regions, freshwater and marine, has not world’s water. This is indeed impressive, but it been adequately tested. should be remembered that most marine taxa are restricted to a relatively narrow photic zone that Freshwater regions hugs coastlines and that marine habitats have been Nearctic (North America) 1060 Neotropical (South and Central America) 8000 much less thoroughly surveyed than those of fresh Palaearctic (Europe, excluding former USSR) 360 water. The little exploration of marine habitats Ethiopian (Africa) 2850 below normal SCUBA depths using submersible or Oriental (Southeast Asia) 3000 unconventional diving techniques suggests that Australian (Australia and New Guinea) 500 there remains considerable unsampled diversity, exemplified by the recent discovery of a second Marine regions Western North Atlantic 1200 species of coelacanth from Sulawesi (Erdmann Mediterranean 400 et al. 1998). Even among deeper-water groups, esti- Tropical western Atlantic 1500 mates of diversity are increasing dramatically. For Eastern North Pacific 600 example, the number of species of Indian Ocean Tropical eastern Pacific 750 grenadiers (Macrouridae; Gadiformes), found at Tropical Indo-West Pacific 4000 depths of 200–2000m, has increased from 30 Temperate Indo-Pacific 2100 known species in 1987 to 114, and is expected to Antarctica 200 far exceed that number (Iwamoto and Anderson 1999). Our brief review of fish distribution is continent-based for freshwater taxa, roughly equivalent to the classic zoogeographic regions of raphy is to make headway. Unfortunately, for Wallace (1876) (Fig. 3.1a), and ocean-based for ma- many areas it is too late to obtain original distri- rine taxa (Table 3.1). The historical reality of these butional data because native faunas have been regions is suspect, with most of them likely to be modified or obliterated through overexploitation, composite, meaning that they are composed of introduction of alien species, or other environ- faunas that have separate evolutionary histories. mental mismanagement (Stiassny 1996; see Other ways of describing zoogeographic regions Reynolds et al., Chapter 15, Volume 2). have been suggested (Fig. 3.1b), but we have no space to describe them. Europe (western Palaearctic)

3.3.1 Freshwater regions The Palaearctic has the most depauperate fresh- water fauna. About 360 species are known in A review of general freshwater fish distribution Europe proper, excluding the former USSR (Kotte- and faunal composition, along with recent and ex- lat 1997). Almost 36% of the diversity is accounted pected developments, can be found in Lundberg for by the Cyprinidae, and another 35% by the et al. (2000). Each section below provides addi- Salmonidae, Coregonidae and Gobiidae combined. tional sources for particular regions. With the ex- Cobitidae, Petromyzontidae, Clupeidae and ceptions of Europe and North America, faunas are Percidae make up an additional 15%, and the re- generally poorly known. Thorough survey and sys- mainder is composed of several poorly represented tematic work is sorely needed if historical biogeog- families (Lundberg et al. 2000). Diversity generally 50 Chapter 3 decreases northwards as recently glaciated regions North America and Mexico (Nearctic region) are encountered. Endemic fishes of international interest include carp (Cyprinus carpio), goldfish The Nearctic fauna is extraordinarily well known, (Carassius auratus) and brown trout (Salmo with over 1060 species of freshwater fishes de- trutta), all widely introduced. Lake Baikal of scribed in North America and Mexico (Lundberg Siberia has a unique fauna derived from the cot- et al. 2000). Although relatively few new species toids (about 50 species), with several endemic are being described, better understanding of genera of Cottidae and two endemic families ecology and genetics with broader adoption of (Abyssocottidae, Comephoridae). The majority of phylogenetic species concepts will undoubtedly the lake’s diversity is found in the Abyssocottidae, add to the total. About 45% of all species belong to with its species generally found below 170m. one of two families, the Cyprinidae (true min- Europe, Asia and North America share many nows) with at least 305 species and the Percidae related taxa, including many fishes. These rela- (perches and darters) with over 170 species. Other tionships have been explored by several workers. dominant families include the Poeciliidae (75 Patterson (1981) compared area cladograms for a species), Catastomidae (68 species), Ictaluridae variety of taxa (Lepisosteidae, osteoglossomorphs, (48 species), Goodeidae (40 species), Fundulidae percopsiforms, Amiidae, Esocoidei, Percidae) and (37 species), Atherinopsidae (35 species), Centrar- found no strongly supported pattern, presenting chidae (32 species), Cottidae (27 species) and Cich- nine alternative cladograms among Europe, Asia lidae (21 species). Over one-third of all species are and eastern and western North America. How- found in the Mississippi Basin, with diversity de- ever, as he pointed out, the fish cladograms avail- clining away from this central drainage area. The able to him were tentative and it would be useful to western Great Basin and Colorado River faunas compare these with cladograms from unrelated have the lowest diversity but exceedingly high en- groups of vertebrates and invertebrates. Mono- demism (over 50%). Northern areas are generally graphic biogeography treatments that investigate depauperate, having been glaciated until only European relationships are uncommon, but in- 10,000 years ago. Fishes of particular interest in- clude the Acipenseriformes (Bemis and Kynard clude Amia calva (Amiidae; only living member of 1997) and Amiidae (Grande and Bemis 1999). the family), Polyodon spathula (with its only liv- The difficulty in tracing biogeographic history of ing relative in China), Hiodon (two species, the Europe, Asia and North America stems from the only living hiodontiforms), Goodeidae (endemic; complicated geological history of Laurasia, the most species have internal fertilization and are northern supercontinent sister to Gondwanaland, euviviparous) and Cyprinodontidae (often restrict- which began a separate history about 180Ma. ed to particular springs). Laurasia’s history involves the transient con- Some of North America’s fishes, such as amiids, nection of Europe and Asia with North America polyodontids and acipenserids, are part of a relict- several times during the Cretaceous and early Ter- ual fauna, with distributions of related taxa that tiary. The intermittent connections explain the date back to the complicated history of Laurasia general similarity and close relationship of these over 100Ma (Early Cretaceous). Others are likely continental faunas, whereas the transient nature to postdate any connections with Eurasia, such as of the connections can provide an explanation for Centrarchidae which is endemic. The Atlantic their more recent divergence. Patterns of internal Gulf Coast was submerged from the late Jurassic to biogeography of the European subcontinent have the Eocene, so freshwater fishes have been in this not been rigorously studied, although glaciation region for no longer than about 40 million years. has undoubtedly had an impact. Individual case The North American fauna has been subjected histories and some comparative work has been to far more rigorous phylogenetic study than fishes done using phylogeographic methods (Bernatchez of any other region. Despite this, very few histori- and Wilson 1998; Bernatchez 2001). cal biogeographic studies that integrate area clado- Historical Biogeography 51 grams from several taxa have been attempted. (Kreiser et al. 2001). If population evolution of An exception is Mayden (1989), who examined postglacial colonizers proves not to be overly re- the cladistic relationships of fishes of the Central ticulate (having a complex pattern of gene ex- Highlands of eastern North America with the aim change), cladistic biogeography techniques would of examining historical influences on their distrib- provide an effective tool to produce area clado- utions. He concluded that relationships among the grams of species and to search for congruence fishes and their distributions were more congruent among species. Congruent area cladograms could with a pre-Pleistocene drainage pattern than with be interpreted as common dispersal routes from the present-day drainages, suggesting a diverse glacial refugia and might form independent tests of and widespread pre-Pleistocene Central Highlands glacial histories provided by geologists. ichthyofauna. This is important for biogeographic A summary of Nearctic biogeography can be studies in North America and other areas peripher- found in Mayden (1992b), although Hocutt and ally affected by glacial periods; distributions of or- Wiley (1986) remains an important reference. ganisms are not necessarily dictated by changes Patterson (1981) summarized area cladograms wrought by glaciation and subsequent dispersal, among northeastern and northwestern North but have a more complex and longer history influ- America with Europe and Asia, but little rigorous enced by ancient distribution patterns as well as study has been undertaken since. Several mono- more modern influences. graphic studies have examined intercontinental However, fish distribution in some parts of biogeography on individual groups (Grande and North America (and Europe) are undoubtedly a Bemis 1991, Polyodontidae; Grande and Bemis product of postglacial dispersal. Much of Canada 1999, Amiidae), although others have been of and parts of the northern USA are likely to broader taxonomic scope (Bemis and Kynard 1997, have been scraped clean of fish fauna during the Acipenseriformes). Wisconsinan glacial period. As the ice sheet re- ceded from its maximum coverage (beginning Africa (Ethiopian region) 18000 years ago), a series of temporary glacial lakes and outlets were formed along its southern Africa is estimated to have 2850 species of fresh- edge. The meltwaters drained into larger basins water fishes (Lundberg et al. 2000), although this and unglaciated drainages, providing dispersal cor- is undoubtedly a considerable underestimate. ridors into the recently deglaciated areas. It is Unlike North American and European faunas, assumed that these colonizing fishes had been new taxa continue to be described at high rates and restricted to waters of unglaciated areas, called distributions are poorly known. Leveque (1997) refugia. Five major refugia have been hypothesized provided a recent review of African ichthyofaunal for North America, with four southern and one provinces. Diversity is highest in tropical areas, northern (Hocutt and Wiley 1986). Studies of as northern and southern regions are quite arid. postglacial dispersal concern themselves with About two-thirds of the diversity is found in seven identifying the refugial origins of a species and families: Cyprinidae (475 species); Characidae and the sequence of glacial waterways used to disperse Citharinidae (208 species); Clariidae, Clarioteidae to their present distributions (e.g. Mandrak and and Mochokidae (345 species); and Cichlidae Crossman 1992). Possible dispersal routes have (conservatively 870 species). Fishes of particular been examined by mapping present-day distribu- interest include several unusual groups such as tions on palaeogeological maps showing glacial the Polypteridae (bichirs), Protopterus (lungfish), outlets and water bodies (Mandrak and Crossman Pantodon (butterflyfish), Mormyridae (elephant- 1992). Phylogeographic methods have been used fishes) and Malapterurus (electric catfishes). The to suggest dispersal routes of individual species diversity of cichlids has attracted most attention, (Wilson and Hebert 1996) or to examine influence with estimates for the three large East African of drainage changes on population structure lakes (Tanganyika, Malawi and Victoria) ranging 52 Chapter 3 over 1000 species (Stiassny and Meyer 1999). selves, but instead form a series of clades that These species flocks exhibit low genetic differenti- often have separate elements in portions of both ation but highly diversified morphologies that continents. evolved independently and rapidly in each lake Special mention should be made of the island of (Fryer and Iles 1972; Meyer 1993). Lake Victoria Madagascar. Once considered to have a depauper- was probably completely dry as recently as 12000 ate fauna, a new accounting of Madagascan fish years ago (Johnson et al. 1996), yet now supports al- species, including the discovery of several new most 400 endemic cichlid species, and the small taxa yet to be described, has put Madagascar more adjoining Lake Nabugabo is only 4000 years old in line with other areas of similar size (Sparks and but has five of seven Haplochromis endemic Stiassny, in press). The total number of freshwater (Greenwood 1981). Lake Tanganyika, which is a species is estimated to be 127, with 76 of these deep rift lake, is believed to have had a succession endemic. Dominant families are the Gobiidae (26 of lowered lake levels of over 600m, providing re- species), Cichlidae (25 species) and Bedotiidae (23 peated isolation events and the opportunity for species). Madagascar possesses a high number speciation. Sympatric speciation via sexual selec- of endemic taxa, many of which appear to occupy tion is also likely to have played a role (Seehausen basal phylogenetic positions (Stiassny and et al. 1999). Raminosa 1994). Despite its geographic proximity With the extremely high level of endemism in to Africa, a preponderance of links with India African waters, approaching 100% at the species and Sri Lanka have been suggested. Its cichlids level and more than 40% at the family level, one (Paratilapia, Ptychochromis), along with south would expect the relationships of the region to Asian genera (Etroplinae), appear to be basal to Europe, Asia and South America to be relatively the African and South/Central American species transparent. However, phylogenetic work on (Stiassny 1991; Farias et al. 1999). The Madagascan African groups is sparse. The recognition of an Pachypanchax (Aplocheilidae) and Ancharias African–South American biogeographic link is not (Ariidae) are also considered basal taxa of their re- new (Eigenmann 1909; Regan 1922), but it has been spective families. Recent work on cichlids (Farias given considerable recent attention (e.g. Lundberg et al. 1999) and aplocheiloids (Murphy and Collins 1993; Grande and Bemis 1999). Many investiga- 1997) suggested that Madagascan and Indian taxa tions take a simple view of area relations, with form a monophyletic sister group to an African and large complex geological entities assumed to South American clade. These patterns are con- behave as single units; essentially two-area gruent with the rifting of the African and South cladograms are examined, with Africa and American landmass from remaining Gondwana- South America as the areas. Indeed, molecular land about 160Ma. The Madagascan genera Bedo- phylogenies of cichlids (Farias et al. 1999) and tia and Rheocles are perhaps most closely related aplocheiloids (killifish) (Murphy and Collins 1997) to eastern Australian melanotaeniids (Aarn and have hypothesized monophyletic African and Ivantsoff 1997), a pattern reflected by possible South American sister clades. These patterns im- close affinities of the Madagascan eleotridid Rat- plicate rifting of Africa and South America from sirakea and Australian/New Guinean Mogurnda. remaining Gondwanaland some 160Ma, and the These relationships and distributions are congru- vicariance of Africa and South America beginning ent with the subsequent separation of Madagascar about 125Ma. However, it would be surprising if and India from the Antarctica–Australian land- each of these continents were not composites and mass about 130Ma. Hence, the freshwater fauna of their relationships not more complicated. Work Madagascar has a very ancient history. by Ortí and Meyer (1997) and Buckup (1998) on characiforms supports this. Unlike cichlids and South and Central America (Neotropical region) killifishes, neither South American nor African characiforms are monophyletic in and of them- The Neotropical region is estimated to have up- Historical Biogeography 53 wards of 8000 species (Vari and Malabarba 1998). South American freshwater fauna. Significant The Amazon basin alone is home to well over 1000 tests of this have yet to be provided. species. The Neotropical fauna is dominated by Biogeographic patterns of South American five major groups: Siluriformes (catfishes, over freshwater fishes are summarized in Vari and 1400 species), Characiformes (characid-like fishes, Malabarba (1998). Available area cladograms are about 1800 species), Gymnotiformes (New World often contradictory or only partially congruent, knifefishes, about 100 species), Cyprinodonti- although this should not be surprising given the formes (rivulines, livebearers and others, about relatively few studies available and the complex 375 species) and Cichlidae (about 450 species) geohistory of the region. The best example in- (Lundberg et al. 2000). Fishes of particular note are volves common patterns among several groups far too numerous to list exhaustively, but include involving northwestern South America and Arapaima gigas (pirarucú, Osteoglossidae, the nearby Central America separated by the Andean largest scaled freshwater fish, reaching 2.5m), Cordilleras. Malabarba (1998) with characiforms some of the smallest and largest catfishes in and Armbruster (1998) with loricariids provided the world (Micromyzon akamai, Aspredinidae, 12 area cladograms that suggested stream capture of mm; Brachyplatystoma, Pimelodidae, 3m), the the Río Tietê (Paraná drainage) by the Río Paraíba electric ‘eel’ (Electrophorus, Gymnotidae, capable in southeastern Brazil during the early Tertiary. of producing up to 650V), and parasitic catfishes or There are also repeated patterns of freshwater in- candirús (Trichomycteridae, famous for reported vasion by marine groups before the completion of entry into the urethra of humans). the Panamanian Isthmus (Nelson 1984, Engrauli- Central America shares the majority of its fau- dae; Lovejoy 1996, Lovejoy et al. 1998, Potamotry- nal elements with northern South America. Poe- gonidae), placing in doubt the oft-cited Pacific ciliids and cichlids are especially well represented origin suggested by Brooks et al. (1981; see Lovejoy in the 350 species known from the region, with 1997). Lovejoy and Collette (2001) used a phy- characids and siluriforms abundant only in the logeny of the Belonidae to suggest four indep- south. The best-known fauna is that of Costa Rica endent entries of fresh water by marine ancestors. through the efforts of Bussing (1998), with 135 In contrast, Dyer (1998) provided an example of species recorded. The most diverse freshwater secondary marine invasion by a freshwater group, families in Costa Rica are Characidae (17), Poecili- sorgentinin silversides; the general assumption of idae (20), Cichlidae (24) and Eleotrididae (16). marine to freshwater invasion clearly requires Central America can boast the most well-known phylogenetic testing. These are only a few exam- application of modern biogeographic methods, ples showing the potential for the explanation of that of Rosen (1978, 1979) and his hypothesis of diversity and faunal composition of South Ameri- congruent taxon and area cladograms for the can fresh waters using historical biogeography. middle American poeciliid genera Heterandria This is an incredibly complex geological system, and Xiphophorus (live bearers). This study was the as shown by Lundberg et al. (1998), that provides an initial trial of Platnick and Nelson’s (1978) cladis- almost endless source of possible processes and tic biogeography and remains one of the convinc- many of the histories will be difficult to tease out. ing examples of geographic congruence for fishes. Careful systematic work and the search for con- Nelson and Ladiges (1996) and Humphries and gruent historical and distributional patterns pro- Parenti (1986, 1999) provide detailed reanalysis vide the best hope for uncovering the processes of Rosen’s data. Although Myers (1938) is asso- by which the unparalleled Neotropical diversity ciated with dispersal scenarios because of Darling- evolved. ton’s (1957) reliance on his ecological fish Relationships of South America to Africa and to divisions, he was open-minded on distribution Australia are briefly explored under those sections. processes, and proposed ‘continental drift to be the Although with somewhat different topologies, ultimate answer’ (Myers 1966, p. 772) for origin of Ortí and Meyer (1997) and Buckup (1998) came to 54 Chapter 3 the conclusion that characiforms had diversified afforded favoured status among at least seven into many of the currently recognized taxa long such lines, even for some marine fishes (Woodland before the separation of South America and Africa 1986) and perhaps invertebrates (Barber et al. about 125Ma. Similarly for siluriforms, de Pinna 2000). Although a zone of biotic contact exists in (1998) has shown that South American taxa are in- the Indonesian area, Parenti (1991, p. 143) logically terspersed in monophyletic clades that include noted that if we cannot place a biogeographic African and sometimes Asian taxa. Much of siluri- boundary to the west or east of a particular island, form diversity existed before the final break-up of perhaps the line should be drawn through it. Ac- Gondwanaland. knowledging that the composite geological nature of several of the region’s islands (Sulawesi, Borneo, New Guinea) is reflected by the evolutionary his- Southern Asia (Oriental region) tory of their taxa will greatly enhance our under- Numbers of freshwater species in the Oriential re- standing of the biogeography of Wallace (see Fig. gion range upwards of 3000, but this is likely an 3.1a). Parenti (1991, 1996) provided area clado- underestimate. Although this has been considered grams based on phallostethids and sicydiine gobi- a single biogeographic unit for almost 150 years, ids and compared these to those derived from bats differences in levels of endemism, diversity and and insects. These and other studies (Michaux faunal composition vary widely from one area 1991, 1996; Parenti 1998) emphasize the compos- to another. Overall, dominant families are the ite nature of many of the Oriental islands and the Cyprinidae (over 1000 species), Balitoridae (300 complex geological history underlying the animal species), Cobitidae (100 species), Bagridae (100 distributions. Hall and Holloway (1998) present species) and Gobiidae (300 species). The continen- several papers outlining details of geology that tal and peninsular portions of the region follow will undoubtedly provide much fodder for biogeo- this faunal composition quite closely, but the graphic processes in the Southeast Asian region. islands (e.g. Philippines, Sulawesi) lose most if Although they provide no fish examples, the ter- not all of the primary freshwater groups and are restrial faunal studies offer phylogenetic patterns dominated by the Gobiidae (Lundberg et al. 2000). to which future fish cladograms and distributions Many of the faunas are poorly known, with India can be compared. and China in particular need of modern systematic work. Noteworthy fishes include Australia and New Guinea (Australian region) the Adrianichthyinae, which carry eggs with the pelvic fins and whose distribution provides some The Australian region has well over 500 fresh- evidence of the geologically composite nature of water fish species. Australia is estimated to have Sulawesi to which it is endemic, and extensive 215 freshwater fish species, with 30 of these having cave fish fauna of the karst regions of China, Laos, been described in the last 10 years; the most recent Vietnam, Thailand, Malaysia and India. review is by Allen (1989). New Guinea has 320 Biogeographic studies of fishes of the region are listed species (Allen 1991) with subsequent re- sparse. India’s drift northwards through the Indian search boosting this to 350. These two major areas Ocean and its effect on its freshwater taxa has share about 50 species (Lundberg et al. 2000). The been debated in the literature (e.g. Briggs 1990; apparent lack of diversity is due in part to substan- Patterson and Owen 1991). The boundary between tial arid areas in Australia, but perhaps more to the traditional Oriental and Australian regions for taxonomic artefact. Many species regarded as terrestrial organisms has been the topic of con- widespread are very likely to comprise several un- siderable debate, with a plethora of boundaries recognized species with more restricted distribu- including and then excluding various islands of tions. As an extreme example, the melanotaeniid the Indonesian archipelago with one or the other Melanotaenia trifasciata has been divided into region (Simpson 1977). Wallace’s Line has been some 35 geographic forms, each with highly re- Historical Biogeography 55 stricted allopatric distributions, suggesting that multiple invasions by marine taxa. An extreme, the number of phylogenetically distinct species and unfounded, example is presented in Aarn and has been grossly underestimated for this and other Ivantsoff (1997) who suggested that the Melano- species in the family. Similar though less dramatic taeniidae, based on the ‘disparate distribution of ‘geographic variation’ has been reported for mem- the family’ in Madagascar, eastern Indonesia, New bers of at least nine additional families. Dominant Guinea and Australia, arose recently from a cos- Australian taxa include the Eleotrididae and Gobi- mopolitan marine atheriniform. Not only would idae (over 50 species), superfamily Galaxioidea, this marine ancestor have had to have gone ex- Melanotaeniidae, Terapontidae and Percichthy- tinct, as there are no basal marine melanotaeniids, idae (each with over 20 species), and Plotosidae and but their hypothesis intimates several indepen- Atherinidae (each with over 15 species). Distinc- dent marine invasions by the ancestor to explain tive components of the fauna include Neocerato- the allopatric distribution of the family. A Gond- dus forsteri (Queensland lungfish), Scleropages (a wanan origin for this group would make a more bonytongue) and Lepidogalaxias salamandroides parsimonious hypothesis (see section on Africa). (salamanderfish), whose relationships have been It has long been acknowledged that the greatly debated (Williams 1997) and whose behav- Australian–New Guinean fauna has a Gondwanan iour includes aestivation (Berra and Pusey 1997). component, particularly through the relationships Unmack (1999, 2001) divided Australia into of galaxioids, Neoceratodus and the osteoglossid several provinces using similarity indices, al- Scleropages. Other taxa, such as the Percichthy- though the historical reality of these needs to be idae and Terapontidae, also appear to have Gond- tested by phylogenetic patterns not yet available. wanan distributions (Lundberg et al. 2000). Until The highest endemism occurs away from the more phylogenies are available, the general pat- eastern and northeastern coastal drainages. In tern will remain elusive and conclusions regarding New Guinea, the most diverse families are the biogeographic relationships unwarranted. Eleotrididae and Gobiidae (about 115 species), Melanotaeniidae (about 70 species), Ariidae (21 3.3.2 Marine regions species), and Terapontidae, Chandidae and Ploto- sidae (each with about 15 species). Melanotaeni- Marine regions are somewhat more fluid and their ids, the rainbowfishes, are well known in the relationships more obscure than freshwater re- aquarium trade. gions. Although some workers have tried to place Australia and New Guinea have been asso- specific boundaries on marine areas, we provide ciated throughout much of their history, made evi- only very general geographic regions because so dent by the shared taxa of these areas. As recently little work has been done to determine their as 6000 years ago during the last glacial sea-level historical reality. Taxonomy, distribution and drop, the two were connected via the sea floor of phylogenies are more poorly known for marine the shallow Arafura Sea and Torres Strait. Streams groups than for freshwater groups. The following of southern New Guinea and adjacent northern is only a brief introduction to some of the marine Australia were confluent at this time. However, areas and is in no way exhaustive or presented as Unmack (1999, 2001) suggested that freshwater a definitive treatment. For example, we refer faunal exchanges were unlikely during this time to McEachran and Fechhelm (1998) for the Gulf due to intervening brackish water and that the of Mexico. The emphasis here is on shorefishes most recent viable freshwater connections likely which are those occurring from shore to a depth of date back to the late Miocene. Unfortunately, phy- 200m; most pelagic groups are widely distributed, logenetic studies are scarce for most freshwater sometimes circumglobal, and deepsea families are fishes of the region, making historical biogeo- usually poorly known both taxonomically and bio- graphic hypotheses sketchy. Many discussions geographically (but see Miya and Nishida 1997 for assume that the freshwater biota resulted from an interesting example). Nelson (1986) discussed 56 Chapter 3 problems and prospects for historical biogeogra- useful in generating testable hypotheses for phy of pelagic fishes. several regions (e.g. Nelson 1984 and White 1986, The application of historical biogeographic amphitropical/isthmian; Springer and Williams techniques to marine fishes was slow to material- 1990, Indo-Pacific). Pattern congruence and phy- ize. Myers’ (1938) observation that primary fresh- logeny will also undoubtedly help to solve the vex- water fishes are confined to particular drainage ing problem of antitropical distributions, where basins and will closely mirror geological history sister species or populations are found in higher held the implication that marine fishes would northern and southern latitudes separated by a have little to say concerning biogeography because broad equatorial distributional gap. of their capacity for dispersal. This dissuaded bio- Some recent ecological work supports the bio- geographers from seriously examining marine geographic evidence for restricted distributions fish distributions for almost half a century. Myers’ and separate histories of some marine fishes. notion of relative value of fishes for biogeography Swearer et al. (1999) and Jones et al. (1999) have continues to have an impact on the choice of study shown, with different experimental protocols, taxa, despite the fact that taxon phylogeny and that despite the production of planktonic eggs and endemism are the final arbiters of biogeographical larvae, offspring of some species remain close to information content rather than little-known their native shores or eventually settle back on physiological tolerances and implied dispersal to the reefs of their parents. The long-distance limitations (e.g. Dyer 1998, p. 528). With the dispersal assumed for marine fishes is not univer- identification of marine zoogeographic regions by sal, and populations of even presently assumed Briggs (1974) and the recognition of areas with high widespread species are unlikely to be panmictic endemism, wide dispersal could not be common to (Cowen et al. 2000; Pogson et al. 2001). Consistent all groups. Springer’s (1982) study of the biogeogra- with these studies, Gill (1999) has questioned phy of shorefishes on the Pacific tectonic plate, marine fish species concepts and the reality of the first major work on marine fishes to espouse widespread species, suggesting that many of a modern approach, was the springboard that these are likely to comprise two or more allopatric launched analytical marine fish biogeography. cryptic species each with more restricted distribu- Hence, analytical marine biogeography is a very tions. These observations have clear taxonomic, new science. Only over the last 15–20 years have systematic, biogeographic and conservation the relatively restricted distributions of several implications. groups, for example blennioids and pseudochro- mids, overcome some of the early hesitancy and North Atlantic and Mediterranean lent credibility to the role of marine shorefishes in biogeographic studies. There are about 1200 species along the North Since Springer (1982), several authors have American Atlantic coast including the northern applied historical biogeographic interpretations Gulf of Mexico. Dominant families are the to the distribution of marine shorefishes (e.g. Serranidae and Gobiidae each with well over 50 White 1986, atherinids; Winterbottom 1986, species, although several families are represented pseudochromids; Springer 1988, Ecsenius, blen- by over 20 species (including Ophichthyidae, niids; Williams 1988, Cirripectes, blenniids; Muraenidae, Gadidae, Syngnathidae, Carangidae, Springer and Williams 1994, Istiblennius, blen- Haemulidae, Sciaenidae, Scorpaenidae). It is niids; Mooi 1995, plesiopids) and even to deep shelf estimated that about 600 species occur in the forms (Harold 1998, sternoptychids). As in fresh- Mediterranean, with almost two-thirds of these water studies, rigorous methods for comparison considered shorefishes. among cladograms has not generally been under- Humphries and Parenti (1999) use data from taken (see Lovejoy 1997 for an exception), but the merluciids (hakes), Lophius (monkfishes) and two search for common distribution patterns has been sponge groups to produce an area cladogram for Historical Biogeography 57

(a) 1

2 1 1 2

Fig. 3.4 Cladistic biogeographic 5 3 analysis of real organisms from 4 3 the North Atlantic. (a) Areas of endemism for chalinid sponges 6 that are similar to those of northern species of the monkfish Lophius 6 genus . 1, Arctic; 2, Boreal; 4 3, Mediterranean; 4, southeastern North Atlantic; 5, western North Atlantic; 6, Caribbean. (b) Taxa and area cladogram of northern (b) (c) Lophius species, with a line drawing of a typical member of the genus. Areas are numbered as in (a). (c) General area cladogram calculated from the combined data of Lophius and two chalinid sponge groups. (Source: after 4 Lophius vaillanti 4 SE North Atlantic Humphries and Parenti 1999, 6 L.gastrophysus 6 Caribbean reproduced by permission of 5 L. americanus 5 Western North Atlantic Oxford University Press. Lophius 3 L. budegassa 3 Mediterranean outline from Nelson 1994; 2 L. piscatorius 2 Boreal reproduced by permission of John 1 L. litulon 1 Arctic Wiley & Sons, Inc.)

the northern and tropical Atlantic (Fig. 3.4). They species for the tropical western Atlantic. This is found that the Mediterranean shared a more recent due in part to differences in defining the area, with faunal history with the Arctic and Boreal regions early workers considering it a single area and later than with any of the more tropical areas. The workers dividing it into the Caribbean including western North Atlantic (eastern USA), then coastal tropical Mexico and Central and South Caribbean and finally the southeastern North America, a West Indian region covering the Atlantic (western Africa, Azores, Canaries and Greater and Lesser Antilles, and a Brazilian region. Cape Verde Islands) are hypothesized as sequen- The latter has recently received attention result- tial outgroups to the Mediterranean and Arctic ing in recognition of additional endemic species regions. Too few corroborative patterns have and better understanding of its faunal history been suggested to imply plausible scenarios. (Joyeux et al. 2001). Böhlke and Chaplin (1993) listed almost 600 species for the Bahamas. A com- prehensive recent accounting for Bermuda (Smith- Tropical western Atlantic Vaniz et al. 1999) lists 365 nearshore species for There are no accurate estimates of numbers of what is recognized to be a relatively depauperate 58 Chapter 3 fauna. Smith-Vaniz et al. (1999) also compares the Eastern North Pacific composition of the Bermudan fauna with that of the Bahamas, Florida Keys and Carolinian Bight, The temperate to boreal areas of the North Pacific comparing a total of over 800 species among these have a surprisingly diverse shorefish fauna. Along areas, which extend as far north as Cape Hatteras. the North American coast alone, from California However, this comparison was among selected to Alaska, there are over 600 species in the families, so a total would be substantially higher top 200m. The dominant families are in the for these areas and the Caribbean as a whole, Scorpaenoidei (sensu Mooi and Gill 1995), with perhaps as high as the 1500 species estimated by scorpaenids (70 species), cottids (85 species) and Lieske and Myers (1996). Dominant families with agonids (over 20 species) making up most of these. 30 or more species each include the Gobiidae, The liparidids comprise about 15 species in water Serranidae, Labrisomidae, Apogonidae and less than 200m deep, the remaining 35 species Sciaenidae, although relatively strong representa- being found at depths between 200 and 7600m. tion is found in the Labridae, Syngnathidae, Other dominant families of the eastern North Carangidae, Chaenopsidae and others with about Pacific include the zoarcids (40–50 species) 20 species each. Endemic families of the western although most are found below 50m, stichaeids Atlantic include the Grammatidae and Inermi- (25 species), pleuronectids (22 species) and em- idae, but there is considerable endemism within biotocids (20 species). The latter family is endemic most represented shorefish families. The to the North Pacific and is unusual in having inter- Caribbean has received substantial attention by nal fertilization and true viviparity. Various other historical biogeographers. Rosen (1976, 1985) pro- taxa suggest a close relationship for North Pacific vides plate tectonic models to explain biotic distri- taxa. For example, additional North Pacific butions. Rauchenberger (1989) and Lydeard et al. endemic families include the gasterosteiform (1995) used freshwater fishes to hypothesize area aulorhynchids, the scorpaenoid Anoplopomatidae relationships, but marine groups have played very (two species) and Hexagrammidae (over a dozen little part in examining Caribbean biogeography. species). Other North Pacific taxa of note are, of Page and Lydeard (1994) summarize the require- course, the Salmonidae with about 11 species, the ments for advancing Caribbean biogeography. most famous of which are anadromous and die That the Caribbean and tropical eastern Pacific after impressive migrations upriver to spawn, and have a close biological connection with many the Pacific halibut (Hippoglossus stenolepis) grow- amphi-American taxa has been known for well ing to over 2.5m and 360kg. over a century. Briggs (1974, 1995) and Rosen (1976) provide summaries from differing perspec- Tropical eastern Pacific tives on this connection. Fishes have provided numerous examples of amphi-American There are about 750 species in the tropical eastern relationships including Atherinopsidae, Dactylo- Pacific, which includes the Gulf of California to scopidae, Labrisomidae, Chaenopsidae and Cen- Ecuador and offshore islands, and over 80% of tropomidae (sensu Mooi and Gill 1995), plus many these are endemic (Allen and Robertson 1994). The others. Sister species and sister taxa at various dominant families are the Sciaenidae (80 species) supraspecific levels are known to be separated by and Gobiidae (80 species), with blennioids also the Isthmus of Panama. Kocher and Stepien (1997) forming a substantial part of the shorefish fauna provide some molecular examples, whereas with Chaenopsidae (30 species), Dactyloscopidae Nelson (1984), White (1986) and Hastings and (24 species) and Labrisomidae (25 species). Other Springer (1994) provide morphological studies well-represented families include the Serranidae, showing relationships among amphi-American Haemulidae and Labridae, all of which have about and pan-Isthmian taxa. 30 species, and Ariidae, Muraenidae and Pomacen- tridae that each have about 20 species. All recorded Historical Biogeography 59 species of the Pomacentridae are endemic. Of the and exposed reefs. The fauna is dominated by offshore islands, the Galapagos has the highest the perciforms, especially the Gobiidae (200+), diversity with 300 species, of which about 17% are Labridae (100+), Pomacentridae (100+), Serrani- endemic; the Clipperton, Cocos, Revillagigedos dae (about 100) and Blenniidae (about 100). and Malpelo islands have around 100 species each Other common and conspicuous families are the with between 5 and 10% endemic. Apogonidae (about 80), Pseudochromidae (about 60), Chaetodontidae (about 50) and Acanthuri- dae (about 45). The only non-perciform families Tropical Indo-Pacific with 50 or more species are the Muraenidae Dividing this huge area, which includes the Red and Syngnathidae. Sea and east coast of Africa to Hawaii and Easter Numbers of species in the Pacific decrease away Island, into identifiable subregions is difficult. from the East Indian region. To the north, Taiwan Essentially all parts of this region require substan- is estimated to have about 2200 species, with al- tial sampling to produce a clear picture of the most 14% of these not found to the south in the fauna. In addition relationships among the fishes East Indian region. The Ryukyu Islands probably remain obscure, and species limits and distribu- have about 2000 species. Almost 1600 species are tions are frequently poorly delimited. Springer expected to be found in Micronesia, but numbers (1982) made an estimate of 4000 species for this decrease eastward through the archipelago (Palau area, but this is likely to be too low. The following 1400, Carolines 1100, Marianas 920, Marshalls discussion of species numbers and endemism fol- and Gilberts 850). Myers (1989) provided an excel- lows Randall (1998) and references therein. lent review of the geography, geology and ecology The highest number of species is found in the of Micronesia and its shorefishes. Springer (1982) East Indian region, which includes Indonesia, New referred to Micronesia as the Caroline Islands con- Guinea and the Philippines, This region is esti- duit, forming a potential set of stepping stones on mated to be home to 2800 species of shorefishes. to the Pacific plate and discussed the biogeography Because the region is so large, it is misleading to of the Pacific in general. The oceanic island groups, provide levels of endemism because many fish are of course, offer fewer habitats and have a less restricted to smaller subregions. The diversity in diverse fauna. However, endemism is often high. this area is probably a result of its complex geologi- Hawaii has about 560 species, with over 130 not cal history. Several islands, including Sulawesi, found elsewhere. The Great Barrier Reef of eastern New Guinea and others, are known to be compo- Australia and New Caledonia have the greatest sites with smaller pieces having accreted during a diversity of the South Pacific, both with about basically counterclockwise rotation of the Sahul 1600 species. Randall et al. (1998) provide the only Shelf into the Sunda Shelf over the last 30 million comprehensive summary of Great Barrier Reef years, with significant tectonic change even over fishes. To the east of New Caledonia, numbers of the last 3–5 million years (Burrett et al. 1991; Hall shore species decrease to about 900 in Samoa, 630 1996, 1998; Moss and Wilson 1998). Significant in the Society Islands, 260 in Rapa and 126 in fluctuations in sea level during glacial cycles are Easter Island (Randall 1998). Highest endemism in also hypothesized to have resulted in speciation. the South Pacific is found at Easter Island, with Shorefishes that appear to provide examples of 22% of its species found nowhere else. The patterns consistent with these geological models Marquesas Islands have 10% endemism, but most are discussed by Winterbottom (1986), Woodland island groups exhibit about 5%. Briggs (1995, (1986), Springer (1988), Springer and Williams 1999a) has attributed these diversity differences (1990, 1994) and Springer and Larson (1996). The to centre of origin hypotheses and competitive region also exhibits an incredible variety of habi- exclusion. However, such explanations have been tats, including silicate and volcanic sands, muds, discredited and replaced by allopatric speciation mangroves, estuaries, lagoons and protected models incorporating plate tectonics, sea-level 60 Chapter 3

fluctuations and other geological mechanisms and torical basis. Particularly striking examples are parameters such as island age (e.g. Springer and antitropical (antiequatorial) distributions, which Williams 1990, 1994). Woodland (1986) suggested are distributions of the same or closely related taxa that Wallace’s Line, usually reserved as a boundary that occur in southern and northern subtropical or for terrestrial taxa (see section on Southern Asia), temperate regions with an intervening equatorial might be descriptive for some marine fish groups, disjunction. Briggs (1999a,b) continued to argue and Barber et al. (2000) have found sharp genetic that such distributions are a result of competitive breaks for stomatopods in this same area. exclusion of newer taxa over older taxa. White West of the East Indian region, numbers of (1989) succinctly discussed the numerous prob- species are surprisingly similar among localities. lems with Briggs’ approach and noted several The Chagos Archipelago has 784 reported species logical alternative hypotheses. As with many but is expected to approach the total of the biogeographic questions, the solution lies in the Maldives at about 950 species (Winterbottom and recognition of pattern. If the antitropical distribu- Anderson 1999). Almost 900 species are reported tions are truly congruent in time and space, then a from the Seychelles, a substantial underestimate common explanation can logically be sought and of the Mauritius shorefish fauna sets its total at tested. However, if each antitropical distribution 670 species, and over 900 species are recorded for is unique, then any explanation of each pattern Oman (Randall 1998). At least 1160 species occur will also be unique and will remain speculative. off the coast of southeast Africa and the Red Sea White (1986, 1989) hypothesized a common has a similar number of species but exhibits a process of global Eocene/Oligocene cooling much higher level of endemism at about 14%, followed by a mid-Miocene equatorial warming although some families have much higher levels, event as the main factor in shaping the repeating for example almost 90% in pseudochromids. patterns of antitropical distributions of taxa, us- Unfortunately, there are no recent summaries ing atherinopsines (silversides) as a model. of the fauna of India, although Anderson (1996) Humphries and Parenti (1986, pp. 84–6) described a estimated over 1000 for Sri Lanka alone. Several novel explanation for antitropical patterns involv- authors have proposed that disjunct distributions ing a north/south division of a pre-Pangaean conti- among sister taxa across the Indian Ocean is a nent called Pacifica. Nelson (1986, p. 216) provided result of vicariance imposed by the movement of an example from the Engraulidae that, if not sup- India northwards towards Asia (Winterbottom porting the Pacifica model, suggested taxonomic 1986; Hocutt 1987; Springer 1988; Mooi 1995). differentiation in a growing rather than shrinking Briggs (1990) criticized these hypotheses based Pacific Basin. Geological explanations that are not mostly on the proposed age of taxa, but Patterson mainstream have been dismissed by some (Hallam and Owen (1991) effectively countered these argu- 1994; Holloway and Hall 1998), but mainstream ments. Within the Indo-Pacific region, the western geologists do not have a stranglehold on truth; Indian Ocean exhibits considerable endemism. plate tectonics itself was dismissed as radical until Several taxa, such as butterflyfish species pairs and recently. pseudochromid genera, have sister-group clades with one member in the western Indian Ocean and Temperate Indo-Pacific (South Africa, the other in the eastern Indian Ocean/western southern Australia, New Zealand) Pacific Ocean (Winterbottom 1986; Gill and Edwards 1999). Although not recognized as a traditional region, this area deserves special mention because of the high diversity and endemism exhibited and its Antitropical distributions faunal connection. In total, this area is home to at Most distributions do not fit classical regions least 2100 shorefishes, with about 900 of these because the regions themselves do not have a his- found nowhere else. South Africa has over 2200 Historical Biogeography 61 species representing over 80% of the world’s fish taxa in common. Cool temperate New Zealand families, with about 13% of species endemic and Australia share over 80 species (in over 40 (Smith and Heemstra 1986). Perhaps 1500 of these families) that are found nowhere else; an addi- are shorefishes, with the highest endemism tional 70 or more species are shared by warm tem- found in the Clinidae (38 endemics), Gobiidae (28 perate environments in these areas (Paulin et al. endemics) and Sparidae (25 endemics). 1989). Families endemic to this region include Ar- Wilson and Allen (1987) estimated that of 600 ripidae (four species) and Chironemidae (four inshore species for southern Australian temperate species). An additional 50 or more species are seas an incredible 85% are endemic, far outstrip- shared with Australia, South Africa and southern ping levels of endemism in any tropical region South America to the exclusion of other regions. (peak of 23% in Hawaii). Families with 20 or more Some of these latter species are from deep water endemics are Clinidae, Gobiesocidae, Gobiidae, and their apparent restricted distributions might Labridae, Monacanthidae and Syngnathidae, be due to poor sampling. However, families such as which includes the famous seadragons. These are the Aplodactylidae are found only in Australia, also the most speciose families, making up over New Zealand, Peru and Chile. Detailed phyloge- one-quarter of inshore species. Other inshore netic studies are necessary to test whether some families with five or more endemic species marine fish distributions might support patterns include Apogonidae, Antennariidae, Atherinidae, shown by freshwater fishes that indicate a shared Callionymidae, Ostraciidae, Platycephalidae, history among Australian, South American and Plesiopidae, Pleuronectidae, Rajidae, Serranidae, African taxa and the implied Gondwanan origin. Scorpaenidae, Soleidae, Tetraodontidae, Urano- scopidae and Urolophidae. Gomon et al. (1994) Antarctic listed 730 species for the southern coast of Australia, although the inclusion of species from The Antarctic Ocean is a well-defined body the pelagic and deepsea realms and geographic of water that is bounded by the continent of coverage were arbitrary. Even from this source, Antarctica to the south and by the Antarctic southern Australia boasts well over 400 endemic convergence, or polar front, which encircles the species representing 100 families and over 60% continent at 50–55° S and is where cold Antarctic endemism. Neira et al. (1998) estimated 1500 surface water (2–6°C) meets warmer Southern temperate Australian species including deepsea Ocean water (8–10°C). Miller (1993) listed almost and oceanic taxa. 300 species for the region, although over one-third Paulin et al. (1989) listed just over 1000 species of these are not shorefishes (they are pelagic or for New Zealand and adjacent islands, with 110 deepwater species). As would be expected, the fish endemic species (11%). Paulin and Roberts (1993) fauna is dominated by those that can withstand ex- split the New Zealand fauna into ‘rockpool’ and ceedingly low temperatures. The Notothenioidei ‘non-rockpool’ fishes and determined that the 58 make up well over one-third of the total fauna and of the 94 species in the former group were endemic over half of the shorefishes, with Nototheniidae whereas 52 of the remaining 900+ fishes were (over 40 species), Artedidraconidae (over 20 endemic. Dominant rockpool endemic groups species), Channichthyidae (about 20 species), are Tripterygiidae (20 endemics), Gobiesocidae (9 Bathydraconidae (16 species), Harpagiferidae endemics), Syngnathidae and Plesiopidae (four (eight species) and Bovichthyidae (one species). endemics each). Of non-rockpool species, domi- Most of these species are endemic to the Antarctic nant endemic families are the Galaxiidae (12 and, given that some species live in water with an endemics; spawns in fresh water but at least part average temperature of about -2°C, exhibit some of the life cycle is marine), Macrouridae (about 12 incredible adaptations: they have special glycopro- endemics) and Pleuronectidae (eight endemics). teins in the blood that lower its freezing point or These localities have a surprising number of lack haemoglobin altogether in the cold oxygen- 62 Chapter 3 rich waters. This suborder is mostly benthic and more on geological hypotheses and environmental without air bladders, but some species have en- and dispersal assumptions (Holloway and Hall tered the water column and have evolved lipid- 1998; Avise 2000); some welcome this shift to a filled sacs that help to provide neutral bouyancy. process/modelling approach (Heaney 1999) where- Bargelloni et al. (2000) hypothesized a phylogeny as others do not (Humphries 2000). Although an based on mitochondrial DNA; a candidate for sis- integration of the two paradigms (process and ter taxon has yet to be established. The Liparidae pattern) will advance the field, process narratives (about 40 species) and Zoarcidae (about 25 species) should be incorporated cautiously. Repeated his- are also common faunal elements but are mostly torical patterns that can form the basis of robust found deeper than 100m, with some deeper than mechanistic explanations are only very slowly 7000m. coming to light but we should be patient in waiting for them to avoid the shortcomings of previous eras in biogeography. 3.4 CONCLUSIONS Historical biogeography in ichthyology is a rela- tively new science and, with few exceptions (e.g. Although biogeographers have been assembling North America), quality data for pattern recogni- taxon and area cladograms for about 25 years, there tion remain scarce; opportunities to make con- have been relatively few successful attempts to tributions to the field abound. Future work in compare and evaluate potential congruence. New biogeography will continue to rely on excellence methods in biogeography are continually devel- in taxonomy and phylogenetics, basic field data for oped, and choosing which advance the field is complete distributional records, and a method difficult. Morrone and Crisci (1995, pp. 392–4) to identify congruence among area cladograms. advocated an integrated approach that can take There are areas of the world for which a synthesis advantage of the merits of each method; such an of phylogenies using rigorous techniques would approach is perhaps as likely to incorporate their help direct future ichthyological work, particu- various disadvantages. Despite the number of larly Neotropical and Australian fresh waters and methods, there is consensus for reliance on the Indo-Pacific. phylogeny reconstruction, use of some form of area cladogram, and increased use of geology and palaeontology. A continuing serious challenge to REFERENCES biogeographers is the definition of the unit of analysis, the area of endemism (Harold and Mooi Aarn and Ivantsoff, W. (1997) Descriptive anatomy of 1994; Morrone 2001). Most workers employ organ- Cairnsichthys rhombosomoides and Iriatherina ismal distributions uncritically without regard to werneri (Teleostei: Atheriniformes), and a phylo- any historical component that underlies the iden- genetic analysis of Melanotaeniidae. Ichthyological tification of homologous areas. Platnick (1991, p. Exploration of Freshwaters 8, 107–50. Allen, G.R. (1989) Freshwater Fishes of Australia. v) states that we should prefer ‘taxa that are maxi- Neptune City, NJ: T.F.H. Publications. mally endemic – those which include the largest Allen, G.R. (1991) Field Guide to the Freshwater Fishes number of species, with the smallest ranges’. He of New Guinea. Madang, Papua New Guinea: Chris- provides a succinct discussion of areas in an intro- tensen Research Institute Publication No. 9, pp. 1–268. duction to a volume that presents a wide range of Allen, G.R. and Robertson, D.R. (1994) Fishes of the applications of the term ‘area’ (Ladiges et al. 1991). Tropical Eastern Atlantic. Honolulu: University of In this chapter, we emphasize the need to iden- Hawaii Press. Anderson, C. (1996) Common Reef Fishes of Sri Lanka. tify patterns of distribution and history before Colombo: The Wildlife Heritage Trust of Sri Lanka. exploring process and mechanisms. There has Arbogast, B. and Kenagy, G.J. (2001) Comparative phylo- been a return to process-orientated approaches, geography as an integrative approach to historical bio- relying less on biological patterns (phylogeny) and geography. Journal of Biogeography 28, 819–25. Historical Biogeography 63

Armbruster, J.W. (1998) Phylogenetic relationships of the Brooks, D.R. and McClennan, D.A. (1991) Phylogeny, suckermouth armored catfishes of the Rhinelepis Ecology, and Behavior: a Research Program in group (Loricariidae: Hypostominae). Copeia 1998, Comparative Biology. Chicago: University of Chicago 620–36. Press. Avise, J.C. (2000) Phylogeography: the History and Brooks, D.R., Thorson, T.B. and Mayes, M.A. (1981) Formation of Species. Cambridge, MA: Harvard Freshwater stingrays (Potamytrogonidae) and their University Press. helminth parasites: testing hypotheses of evolution Barber, P.H., Palumbi, S.R., Erdmann, M.V. and Moosa, and coevolution. In: V.A. Funk and D.R. Brooks (eds) M.K. (2000) A marine Wallace’s line? Nature 406, Advances in Cladistics: Proceedings of the First Meet- 692–3. ing of the Willi Hennig Society. New York: New York Bargelloni, L., Marcato, S., Zane, L. and Patarnello, T. Botanical Garden, pp. 147–75. (2000) Mitochondrial phylogeny of notothenioids: a Brown, J.H. and Lomolino, M.V. (1998) Biogeography, molecular approach to Antarctic fish evolution and 2nd edn. Sunderland, MA: Sinauer. biogeography. Systematic Biology 49, 114–29. Buckup, P.A. (1998) Relationships of the Characidiinae Bemis, W.E. and Kynard, B. (1997) Sturgeon rivers: an and phylogeny of characiform fishes (Teleostei: introduction to acipenseriform biogeography and Ostariophysi). In: L.R. Malabarba, R.E. Reis, R.P. life history. Environmental Biology of Fishes 48, Vari, Z.M. Lucena and C.A.S. Lucena (eds) Phylogeny 167–83. and Classification of Neotropical Fishes. Porto Bernatchez, L. (2001) The evolutionary history of brown Alegre, Brazil: Edipucrs, pp. 123–44. trout (Salmo trutta L.) inferred from phylogeographic, Burrett, C., Duhig, N., Berry, R. and Varne, R. (1991) nested clade, and mismatch analyses of mitochondrial Asian and south-western Pacific continental terranes DNA variation. Evolution 55, 351–79. derived from Gondwana, and their biogeographic Bernatchez, L. and Wilson, C.C. (1998) Comparative significance. Australian Systematic Botany 4, phylogeography of Nearctic and Palearctic freshwater 13–24. fishes. Molecular Ecology 7, 431–52. Bussing, W.A. (1998) Freshwater fishes of Costa Rica, 2nd Berra, T.M. and Pusey, B.J. (1997) Threatened fishes of the edn. San José: Editorial de la Universidad de Costa world: Lepidogalaxias salamandroides Mees, 1961 Rica. (Lepidogalaxiidae). Environmental Biology of Fishes Cowen, R.K., Lwiza, K.M.M., Sponaugle, S., Paris, C.B. 50, 201–2. and Olson, D.B. (2000) Connectivity of marine popula- Böhlke, J.E. and Chaplin, C.C.G. (1993) Fishes of the tions: open or closed? Science 287, 857–9. Bahamas and Adjacent Tropical Waters, 2nd edn. Craw, R.C., Grehan, J.R. and Heads, M.J. (1999) Panbio- Austin: University of Texas Press. geography: Tracking the History of Life. New York: Bowen, B.W. and Grant, W.S. (1997) Phylogeography of Oxford University Press. the sardines (Sardinops spp.): assessing biogeographic Croizat, L. (1964) Space, Time, Form: the Biological models and population histories in temperate Synthesis. Caracas: published by the author. upwelling zones. Evolution 51, 1601–10. Darlington, P.J. Jr (1957) Zoogeography: the Geographi- Bowen, B.W., Bass, A.L., Rocha, L.A., Grant, W.S. and cal Distibution of Animals. New York: John Wiley & Robertson, D.R. (2001) Phylogeography of the trum- Sons. petfishes (Aulostomus): ring species complex on a de Pinna, M.C.C. (1998) Phylogenetic relationships of global scale. Evolution 55, 1029–39. Neotropical Siluriformes (Teleostei: Ostariophysi): Bremer, K. (1992) Ancestral areas: a cladistic reinterpre- historical overview and synthesis of hypotheses. In: tation of the center of origin concept. Systematic Biol- L.R. Malabarba, R.E. Reis, R.P. Vari, Z.M. Lucena and ogy 41, 436–45. C.A.S. Lucena (eds) Phylogeny and Classification of Briggs, J.C. (1974) Marine Zoogeography. New York: Neotropical Fishes. Porto Alegre, Brazil: Edipucrs, McGraw-Hill. pp. 279–330. Briggs, J.C. (1990) The historic biogeography of India: iso- Dyer, B.S. (1998) Phylogenetic systematics and historical lation or contact? Systematic Zoology 38, 322–32. biogeography of the Neotropical silverside family Briggs, J.C. (1995) Global Biogeography. Amsterdam: Atherinopsidae (Teleostei: Atheriniformes). In: L.R. Elsevier. Malabarba, R.E. Reis, R.P. Vari, Z.M. Lucena and Briggs, J.C. (1999a) Extinction and replacement in the C.A.S. Lucena (eds) Phylogeny and Classification of Indo-West Pacific Ocean. Journal of Biogeography 26, Neotropical Fishes. Porto Alegre, Brazil: Edipucrs, 777–83. pp. 519–36. Briggs, J.C. (1999b) Coincident biogeographic patterns: Eigenmann, C.H. (1909) The freshwater fishes of Indo-West Pacific Ocean. Evolution 53, 326–35. Patagonia and an examination of the Archiplata- 64 Chapter 3

Archhelenis theory. Reports of the Princeton Univer- Southeast Asia. Geological Society Special Publica- sity Expeditions to Patagonia, Zoology 3, 225–374. tion No. 106, pp. 153–84. Erdmann, M.V., Caldwell, R.L. and Moosa, M.K. (1998) Hall, R. (1998) The plate tectonics of Cenozoic SE Asia Indonesian ‘King of the Sea’ discovered. Nature 395, and the distribution of land and sea. In: R. Hall and J.D. 335. Holloway (eds) Biogeography and Geological Evolu- Eschmeyer, W.N. (1998) Catalog of Fishes. San Francisco: tion of SE Asia. Leiden: Backhuys Publishers, pp. California Academy of Sciences. 99–131. Farias, I.P., Ortì, G., Sampaio, I., Schneider, H. and Meyer, Hall, R. and J.D. Holloway (eds) (1998) Biogeography and A. (1999) Mitochondrial DNA phylogeny of the family Geological Evolution of SE Asia. Leiden: Backhays Cichlidae: monophyly and fast molecular evolution Publishers. of the neotropical assemblage. Journal of Molecular Hallam, A. (1994) An Outline of Phanerozoic Biogeogra- Evolution 48, 703–11. phy. Oxford: Oxford University Press. Farris, J.S. (1981) Distance data in phylogenetic analysis. Harold, A.S. (1998) Areas of endemism and historical in- In: V.A. Funk and D.R. Brooks (eds) Advances in ference in biogeography. In: A.C. Pierrot-Bults, R.K. Cladistics: Proceedings of the First Meeting of the Johnson, S. van der Spoel and B.J. Zahuranec (eds) Willi Hennig Society. New York: New York Botanical Pelagic Biogeography: Proceedings of the Second Garden, pp. 3–23. International Conference on Pelagic Biogeography, Forbes, E. (1856) Map of the distribution of marine life. In: Noordwijkerhout, The Netherlands, 9–14 July, 1995. A.K. Johnston (ed.) The Physical Atlas of Natural UNESCO Intergovernmental Oceanographic Com- Phenomena. Edinburgh and London: W. and A.K. mission Workshop Report No. 142, pp. 148–55. Johnston, Plate No. 31. Harold, A.S. and Mooi, R.D. (1994) Areas of endemism: Fryer, G. and Iles, T.D. (1972) The Cichlid Fishes of the definition and criteria for recognition. Systematic Great Lakes of Africa. Edinburgh: Oliver and Boyd. Biology 43, 261–6. Gill, A.C. (1999) Subspecies, geographic forms and wide- Hastings, P.A. and Springer, V.G. (1994) Review of Stath- spread Indo-Pacific coral-reef fish species: a call for monotus, with redefinition and phylogenetic analysis change in taxonomic practice. In: B. Seret and J.-Y. of the Chaenopsidae (Teleostei: Blennioidei). Smith- Sire (eds) Proceedings of the Fifth Indo-Pacific Fish sonian Contributions to Zoology 558, 1–48. Conference, Nouméa, 1997. Paris: Societé Français Heaney, L.R. (1999) Historical biogeography in S.E. Asia: Ichtyologie, pp. 79–87. integrating paradigms and refining the details. Journal Gill, A.C. and Edwards, A.J. (1999) Monophyly, inter- of Biogeography 26, 435–7. relationships and description of three new genera in Hocutt, C.H. (1987) Evolution of the Indian Ocean and the dottyback fish subfamily Pseudoplesiopinae the drift of India. Hydrobiologia 150, 203–23. (Teleostei: Perciformes: Pseudochromidae). Records of Hocutt, C.H. and Wiley, E.O. (1986) The Zoogeography of the Australian Museum 51, 141–60. North American Freshwater Fishes. New York: John Gomon, M.F., Glover, J.C.M. and Kuiter, R.H. (1994) The Wiley and Sons. Fishes of Australia’s South Coast. Adelaide: State Holloway, J.D. and Hall, R. (1998) SE Asian geology and Print. biogeography: an introduction. In: R. Hall and J.D. Grande, L. and Bemis, W.E. (1991) Osteology and phylo- Holloway (eds) Biogeography and Geological genetic relationships of fossil and Recent paddlefishes Evolution of SE Asia. Leiden: Backhuys Publishers, pp. (Polyodontidae) with comments on the interrelation- 1–23. ships of Acipenseriformes. Journal of Vertebrate Humphries, C.J. (1989) Fundamental biogeography. Paleontology Supplement 11, 1–121. Journal of Biogeography 16, 401–2. Grande, L. and Bemis, W.E. (1999) Historical biogeogra- Humphries, C.J. (2000) Form, space and time: which phy and historical paleoecology of Amiidae and other comes first? Journal of Biogeography 27, 11–15. halecomorph fishes. In: G. Arratia and H.-P. Schultze Humphries, C.J. and Parenti, L.R. (1986) Cladistic (eds) Mesozoic Fishes 2: Systematics and Fossil Biogeography. Oxford: Clarendon Press. Record. München, Germany: Verlag, pp. 413–24. Humphries, C.J. and Parenti, L.R. (1999) Cladistic Greenwood, P.H. (1981) Species flocks and explosive Biogeography, 2nd edn. Oxford: Clarendon Press. speciation. In: P.H. Greenwood and P.L. Forey (eds) Iwamoto, T. and Anderson, M.E. (1999) New species of Chance, Change and Challenge: the Evolving Caelorinchus (Macrouridae, Gadiformes, Teleostei) Biosphere. London: Cambridge University Press, from the Indian Ocean. Copeia 1999, 1079–83. pp. 61–74. Johnson, T.C., Scholz, C.A., Talbot, M.R. et al. (1996) Hall, R. (1996) Reconstructing Cenozoic SE Asia. In: Late Pleistocene desiccation of Lake Victoria and rapid R. Hall and D. Blundell (eds) Tectonic Evolution of evolution of cichlid fishes. Science 273, 1091–3. Historical Biogeography 65

Jones, G.P., Milicich, M.J., Emslie, M.J. and Lunow, C. (eds) Phylogeny and Classification of Neotropical (1999) Self-recruitment in a coral reef fish population. Fishes. Porto Alegre, Brazil: Edipucrs, pp. 13–48. Nature 402, 802–4. Lundberg, J.G., Kottelat, M., Smith, G.R., Stiassny, Joyeux, J.-C., Floeter, S.R., Ferreira, C.E.L. and Gasparini, M.L.J. and Gill, A.C. (2000) So many fishes, so little J.L. (2001) Biogeography of tropical reef fishes: the time: an overview of recent ichthyological discovery in South Atlantic puzzle. Journal of Biogeography 28, continental waters. Annals of the Missouri Botanical 831–41. Garden 87, 26–62. Kocher, T.D. and Stepien, C.A. (1997) Molecular System- Lydeard, C., Wooten, M.C. and Meyer, A. (1995) Mole- atics of Fishes. San Diego, CA: Academic Press. cules, morphology, and area cladograms: a cladistic Kottelat, M. (1997) European freshwater fishes. An and biogeographic analysis of Gambusia (Teleostei: heuristic checklist of the freshwater fishes of Europe Poeciliidae). Systematic Biology 44, 221–36. (exclusive of former USSR), with an introduction for Malabarba, L.R. (1998) Monophyly of the Cheirodon- non-systematists and comments on nomenclature and tinae, characters and major clades (Ostariophysi: conservation. Biologia (Bratislava), Section Zoology Characidae). In: L.R. Malabarba, R.E. Reis, R.P. Vari, 52 (Suppl. 5), 1–271. Z.M. Lucena and C.A.S. Lucena (eds) Phylogeny and Kreiser, B., Mitton, J. and Woodling, J. (2001) Phylo- Classification of Neotropical Fishes. Porto Alegre, geography of the plains killifish, Fundulus zebrinus. Brazil: Edipucrs, pp. 193–233. Evolution 55, 339–50. Mandrak, N.E. and Crossman, E.J. (1992) Postglacial Ladiges, P.Y., Humphries, C.J. and Martinelli, L.W. (eds) dispersal of freshwater fishes into Ontario. Canadian (1991) Austral Biogeography. Melbourne: CSIRO Journal of Zoology 70, 2247–59. Editorial Services Section. Mayden, R.L. (1989) Vicariance biogeography, par- Leveque, C. (1997) Biodiversity Dynamics and Conser- simony, and evolution in North American freshwater vation of Freshwater Fishes of Tropical Africa. Cam- fishes. Systematic Zoology 37, 329–55. bridge: Cambridge University Press. Mayden, R.L. (1992a) The wilderness of panbiogeogra- Lieske, E. and Myers, R. (1996) Coral Reef Fishes. Prince- phy: a synthesis of space, time, and form? (A review ton, NJ: Princeton University Press. of) Panbiogeography Special Issue, New Zealand Lovejoy, N.R. (1996) Systematics of myliobatid elasmo- Journal of Zoology 16(4). Systematic Zoology 40, branchs: with emphasis on the phylogeny and histori- 503–19. cal biogeography of neotropical freshwater stingrays Mayden, R.L. (1992b) Systematics, Historical Ecology, (Potamotrygonidae: Rajiformes). Zoological Journal of and North American Freshwater Fishes. Stanford, CA: the Linnean Society 117, 207–57. Stanford University Press. Lovejoy, N.R. (1997) Stingrays, parasites, and neotropical Mayr, E. (1982) The Growth of Biological Thought: biogeography: a closer look at Brooks et al.’s hypoth- Diversity, Evolution and Inheritance. Cambridge, eses concerning the origins of neotropical freshwater MA: Belknap Press. rays (Potamotrygonidae). Systematic Biology 46, McEachran, J.D. and Fechhelm, J.D. (1998) Fishes of 218–30. the Gulf of Mexico: Volume 1 Myxiniformes to Lovejoy, N.R. and Collette, B.B. (2001) Phylogenetic Gasterosteiformes. Austin: University of Texas Press. relationships of New World needlefishes (Teleostei: Meyer, A. (1993) Phylogenetic relationships and evolu- Belonidae) and the biogeography of transitions be- tionary processes in East African cichlid fishes. Trends tween marine and freshwater habitats. Copeia 2001, in Ecology and Evolution 8, 279–84. 324–38. Michaux, B. (1991) Distributional patterns and tectonic Lovejoy, N.R., Bermingham, E. and Martin, A.P. (1998) development in Indonesia: Wallace reinterpreted. Marine incursion into South America. Nature 396, Australian Systematic Botany 4, 25–36. 421–2. Michaux, B. (1996) The origin of southwest Sulawesi and Lundberg, J.G. (1993) African–South American fresh- other Indonesian terranes: a biological view. Palaeo- water fish clades and continental drift: problems with a geography: Palaeoclimatology and Palaeoecology paradigm. In: P. Goldblatt (ed.) Biological Relation- 122, 167–83. ships between Africa and South America. New Haven, Miller, R.G. (1993) A History and Atlas of the Fishes CT: Yale University Press, pp. 156–99. of the Antarctic Ocean. Carson City, NV: Foresta Lundberg, J.G., Marshall, L.G., Guerrero, J., Horton, B., Institute. Malabarba, M.C.S.L. and Wesselingh, F. (1998) The Miya, M. and Nishida, M. (1997) Speciation in the open stage for Neotropical fish diversification: a history of ocean. Nature 389, 803–4. tropical South American rivers. In: L.R. Malabarba, Mooi, R.D. (1995) Revision, phylogeny, and discussion of R.E. Reis, R.P. Vari, Z.M. Lucena and C.A.S. Lucena biology and biogeography of the fish genus Plesiops 66 Chapter 3

(Perciformes: Plesiopidae). Royal Ontario Museum biogeography and analysis of paralogy-free subtrees. Life Sciences Contributions 159, 1–108. American Museum Novitates No. 3167, 1–58. Mooi, R.D. and Gill, A.C. (1995) Association of epaxial Nelson, G. and Platnick, N. (1981) Systematics and musculature with dorsal-fin pterygiophores in acan- Biogeography: Cladistics and Vicariance. New York: thomorph fishes, and its phylogenetic significance. Columbia University Press. Bulletin of the Natural History Museum, London, Nelson, G. and Platnick, N. (1984) Biogeography. Zoology Series, 61, 121–37. Carolina Biology Reader 119, 1–16. Burlington, NC: Morrone, J.J. (2001) Homology, biogeography and areas Carolina Biological Supply Co. of endemism. Diversity and Distributions 7, Nelson, G. and Rosen, D.E. (1981) Vicariance Biogeogra- 297–300. phy: a Critique. New York: Columbia University Morrone, J.J. and Crisci, J.V. (1995) Historical biogeogra- Press. phy: introduction to methods. Annual Review of Nelson, J.S. (1994) Fishes of the World, 3rd edn. New Ecology and Systematics 26, 373–401. York: John Wiley & Sons, p. 241. Moss, S.J. and Wilson, M.E.J. (1998) Biogeographic impli- Ortí, G. and Meyer, A. (1997) The radiation of characi- cations from the Tertiary palaeogeographic evolution form fishes and the limits of resolution of mitochon- of Sulawesi and Borneo. In: R. Hall and J.D. Holloway drial ribosomal DNA sequences. Systematic Biology (eds) Biogeography and Geological Evolution of SE 46, 75–100. Asia. Leiden: Backhuys Publishers, pp. 133–63. Page, R.D.M. (1989) New Zealand and the new biogeogra- Murphy, R.W. and Doyle, K.D. (1998) Phylophenetics: phy. New Zealand Journal of Zoology 16, 471–83. frequencies and polymorphic characters in genealogi- Page, R.D.M. and Lydeard, C. (1994) Towards a cladistic cal estimation. Systematic Biology 47, 737–61. biogeography of the Caribbean. Cladistics 10, 21–41. Murphy, W.J. and Collins, G.E. (1997) A molecular Parenti, L.R. (1991) Ocean basins and the biogeography of phylogeny for aplocheiloid fishes (Atherinomorpha, freshwater fishes. Australian Systematic Botany 4, Cyprinodontiformes): the role of vicariance and the 137–49. origin of annualism. Molecular Biology and Evolution Parenti, L.R. (1996) Phylogenetic systematics and bio- 14, 790–9. geography of phallostethid fishes (Atherinomorpha, Myers, A.A. and Giller, P.S. (1988) Analytical Bio- Phallostethidae) of northwestern Borneo, with de- geography: An Integrated Approach to the Study of scription of a new species. Copeia 1996, 703–12. Animal and Plant Distributions. London: Chapman & Parenti, L.R. (1998) Neostethus djajaorum, new species, Hall. from Sulawesi, Indonesia, the first phallostethid Myers, G.S. (1938) Fresh-water fishes and West Indian fish (Teleostei: Atherinomorpha) known from east of zoogeography. Annual Report of the Smithsonian Wallace’s Line. The Raffles Bulletin of Zoology 46, Institution for 1937 3465, 339–64. 139–50. Myers, G.S. (1966) Derivation of the freshwater fauna of Patterson, C. (1981) The development of the North Central America. Copeia 1966, 766–73. American fish fauna: a problem of historical biogeogra- Myers, R.F. (1989) Micronesian Reef Fishes. Guam: Coral phy. In: P.L. Forey (ed.) The Evolving Biosphere: Graphics. Chance, Change, Challenge. Cambridge: British Neira, F.J., Miskiewicz, A.G. and Trnski, T. (1998) Larvae Museum (Natural History) and Cambridge University of Temperate Australian Fishes: Laboratory Guide for Press, pp. 265–81. Larval Fish Identification. Nedlands, WA: University Patterson, C. and Owen, H.G. (1991) Indian isolation or of Western Australia Press. contact? A response to Briggs. Systematic Zoology 40, Nelson, G. (1978) From Candolle to Croizat: comments 96–100. on the history of biogeography. Journal of Historical Paulin, C. and Roberts, C. (1993) Biogeography of New Biology 11, 269–305. Zealand rockpool fishes. In: C.N. Battershill, D.R. Nelson, G. (1984) Identity of the anchovy Hilde- Schiel, G.P. Jones, R.G. Creese and A.B. MacDiarmid brandichthys setiger with notes on relationships (eds) Proceedings of the Second International and biogeography of the genera Engraulis and Temperate Reef Symposium, 7–10 January 1992. Cetengraulis. Copeia 1984, 422–7. Wellington, New Zealand: National Institute of Water Nelson, G. (1986) Models and prospects of historical bio- and Atmospheric Research Marine, pp. 191–9. geography. In: A.C. Pierrot-Bults, S. van der Spoel, B.J. Paulin, C., Stewart, A., Roberts, C. and McMillan, P. Zahuranec and R.K. Johnson (eds) Pelagic Biogeogra- (1989) New Zealand Fish: a Complete Guide. Welling- phy. UNESCO Technical Papers in Marine Science ton: Government Printing Office. Vol. 49, pp. 214–18. Platnick, N.I. (1991) Commentary: on areas of en- Nelson, G. and Ladiges, P.Y. (1996) Paralogy in cladistic demism. In: P.Y. Ladiges, C.J. Humphries and L.W. Historical Biogeography 67

Martinelli (eds) Austral Biogeography. Melbourne: Smith-Vaniz, W.F., Collette, B.B. and Luckhurst, B.E. CSIRO Editorial Services Section, pp. v–vi. (1999) Fishes of Bermuda: History, Zoogeography, Platnick, N. and Nelson, G. (1978) A method of analysis Annotated Checklist, and Identification Keys. in historical biogeography. Systematic Zoology 27, Lawrence, KS: Allen Press, American Society of 1–16. Ichthyologists and Herpetologists Special Publication Platnick, N. and Nelson, G. (1989) Spanning-tree bio- No. 4. geography: shortcut, detour, or dead-end? Systematic Sparks, J.S. and Stiassny, M.L.J. (in press) Introduction to Zoology 37, 410–19. the freshwater fishes. In: S.M. Goodman and J.P. Pogson, G.H., Taggart, C.T., Mesa, K.A. and Boutilier, Benstead (eds) The Natural History of Madagascar. R.G. (2001) Isolation by distance in the Atlantic cod, Chicago: Chicago: University Press. Gadus morhua, at large and small geographic scales. Springer, V.G. (1982) Pacific plate biogeography, with Evolution 55, 131–46. special reference to shorefishes. Smithsonian Contri- Randall, J.E. (1998) Zoogeography of shore fishes of the butions to Zoology 367, 1–182. Indo-Pacific region. Zoological Studies 37, 227–68. Springer, V.G. (1988) The Indo-Pacific blenniid fish genus Randall, J.E., Allen, G.R. and Steene, R.C. (1998) Fishes Ecsenius. Smithsonian Contributions to Zoology 465, of the Great Barrier Reef and Coral Sea. Honolulu: 1–134. University of Hawaii Press. Springer, V.G. and Larson, H.K. (1996) Pholidichthys Rauchenberger, M. (1989) Historical biogeography of anguis, a new species of pholidichthyid fish from poeciliid fishes in the Caribbean. Systematic Zoology Northern Territory and Western Australia. Proceed- 37, 356–65. ings of the Biological Society of Washington 109, Regan, C.T. (1922) The distribution of the fishes of 353–65. the order Ostariophysi. Bijdragen Tot de Dierkunde, Springer, V.G. and Williams, J.T. (1990) Widely distrib- Leiden 22, 203–7. uted Pacific Plate endemics and lowered sea-level. Rosen, B.R. (1984) Reef coral biogeography and climate Bulletin of Marine Science 47, 631–40. through the late Cainozoic: just islands in the sun or a Springer, V.G. and Williams, J.T. (1994) The Indo-West critical pattern of islands. Geological Journal Special Pacific blenniid fish genus Istiblennius, Blenniella, Issue 11, 201–56. and Paralticus, new genus. Smithsonian Contribu- Rosen, D.E. (1976) A vicariance model of Caribbean bio- tions to Zoology 565, 1–193. geography. Systematic Zoology 24, 431–64. Stiassny, M.L.J. (1991) Phylogenetic intrarelationships of Rosen, D.E. (1978) Vicariant patterns and historical the family Cichlidae: an overview. In: M.H.A. Keenley- explanation in biogeography. Systematic Zoology 27, side (ed.) Cichlid Fishes: Behaviour, Ecology and 159–88. Evolution. London: Chapman & Hall, pp. 1–35. Rosen, D.E. (1979) Fishes from the uplands and inter- Stiassny, M.L.J. (1996) An overview of freshwater bio- montane basins of Guatemala: revisionary studies and diversity: with some lessons from African fishes. comparative geography. Bulletin of the American Fisheries 21, 7–13. Museum of Natural History 162, 269–375. Stiassny, M.L.J. and Meyer, A. (1999) Cichlids of the rift Rosen, D.E. (1985) Geological hierarchies and biogeo- lakes. Scientific American 280, 64–9. graphic congruence. Annals of the Missouri Botanical Stiassny, M.L.J. and Raminosa, N. (1994) The fishes of the Garden 72, 636–59. inland waters of Madagascar. Biological diversity in Sclater, P.L. (1858) On the general geographical distribu- African fresh- and brackish water fishes. Geographical tion of the members of the class Aves. Journal of the overviews. Annales du Musée Royal de l’Afrique Linnean Society of London, Zoology 2, 130–45. Centrale, Zoology 275, 133–49. Seberg, O. (1986) A critique of the theory and methods of Swearer, S.E., Caselle, J.E., Lea, D.W. and Warner, R.R. panbiogeography. Systematic Zoology 35, 369–80. (1999) Larval retention and recruitment in an island Seehausen, O., van Alphen, J.J.M. and Lande, R. (1999) population of a coral-reef fish. Nature 402, 799–802. Color polymorphism and sex ratio distortion in a cich- Tassy, P. and Deleporte, P. (1999) Hennig 17, Sao Paulo lid fish as an incipient stage in sympatric speciation by 21–25 September 1998. Bulletin de la societé sexual selection. Ecology Letters 2, 367–78. Française de Systematique, 21 Janvier 1999, 13–14. Simpson, G.G. (1977) Too many lines: the limits of Unmack, P. (1999) Biogeography of Australian fresh- the Oriental and Australian zoogeographic regions. water fishes. Unpublished MSc thesis, Arizona State Proceedings of the American Philosophical Society University. 121, 107–20. Unmack, P.J. (2001) Biogeography of Australian fresh- Smith, M.M. and Heemstra, P.C. (1986) Smiths’ Sea water fishes. Journal of Biogeography 28, 1053–89. Fishes. Johannesburg: Macmillan South Africa. Vari, R.P. and Malabarba, L.R. (1998) Neotropical 68 Chapter 3

ichthyology: an overview. In: L.R. Malabarba, R.E. and D.W. Walton (eds) Fauna of Australia, Vol. 1A. Reis, R.P. Vari, Z.M. Lucena and C.A.S. Lucena (eds) Canberra: Australian Government Publishing Phylogeny and Classification of Neotropical Fishes. Service, pp. 43–68. Porto Alegre, Brazil: Edipucrs, pp. 1–11. Wilson, C.C. and Hebert, P.D.N. (1996) Phylogeographic Wallace, A.R. (1876) The Geographical Distribution of origins of lake trout (Salvelinus namaycush) in eastern Animals. London: Macmillan. North America. Canadian Journal of Fisheries and White, B.N. (1986) The Isthmian link, antitropicality and Aquatic Sciences 53, 2764–75. American biogeography: distributional history of the Winterbottom, R. (1986) Revision and vicariance bio- Atherinopsinae (Pisces: Atherinidae). Systematic geography of the subfamily Congrogadinae (Pisces: Zoology 35, 176–94. Perciformes: Pseudochromidae). Indo-Pacific Fishes 9, White, B.N. (1989) Antitropicality and vicariance: a reply 1–34. to Briggs. Systematic Zoology 38, 77–9. Winterbottom, R. and Anderson, R.C. (1999) Fishes of the Williams, J.T. (1988) Revision and phylogenetic relation- Chagos Archipelago. In: C.R.C. Sheppard and M.R.D. ships of the blenniid fish genus Cirripectes. Indo- Seaward (eds) Ecology of the Chagos Archipelago. Pacific Fishes 17, 1–78. London: Linnean Society Occasional Publications, Williams, R.R.G. (1997) Bones and muscles of the sus- pp. 101–17. pensorium in the galaxioids and Lepidogalaxias Woodland, D.J. (1986) Wallace’s Line and the distribution salamandroides (Teleostei: Osmeriformes) and their of marine inshore fishes. In: T. Uyeno, R. Arai, T. phylogenetic significance. Records of the Australian Taniuchi and K. Matsuura (eds) Indo-Pacific Fish Museum 49, 139–66. Biology: Proceedings of the Second International Con- Wilson, B.R. and Allen, G.R. (1987) Marine components ference on Indo-Pacific Fishes. Tokyo: Ichthyological and distribution of marine fauna. In: G.R. Dyne Society of Japan, pp. 453–60. Part 2 Production and Population Structure 4 The Physiology of Living in Water

OLE BRIX

4.1 INTRODUCTION especially the river drainages of Southeast Asia and South America. About 500 marine species About 25000 living species, which include more enter fresh water during their life cycle (see Gill than 50% of all living vertebrate species, belong and Mooi, Chapter 2, Mooi and Gill, Chapter 3 and to a very heterogeneous assemblage of aquatic Metcalfe et al., Chapter 8, this volume). animals called fish (Nelson 1994; Gill and Mooi, The diversity of adaptation to these habitats Chapter 2, this volume). They have been divided is determined by the evolutionary background of into three major groupings that have been the fish and the physical and chemical charac- separated during the last 500 million years of teristics of water. These characteristics set a evolution. The Agnatha, or jawless vertebrates, are number of constraints on the functional design of primitive fish that evolved in sea water about fish, of which high density, low compressibility, 350–500 million years ago. The ostracoderms, also solvent properties and transparency are the most an ancient lineage, are largely found in freshwater important. deposits. This line is today represented by the 1 The high density of water reduces the effects hagfishes (about 43 species) and lampreys (about of gravity and enables fish to remain suspended in 41 species). A major group of fishes includes the the water column using a minimum amount of Chondrichthyes or cartilaginous fishes, with 800+ energy. Fish cope with buoyancy problems by types of rays and sharks and about 50 or so species applying either dynamic or static lift or a combina- of chimeras, and the Osteichthyes, which includes tion of both. the bony or teleost fishes and contains more than 2 The low compressibility of water strongly in- 20000 species. The cartilaginous fishes evolved fluences the swimming performance of fish, who like the Agnatha in sea water. The teleost fish have to push large volumes of water aside to move evolved in fresh water 50–100 million years ago. through it. For this purpose they have developed They proliferated into a huge number of species different motor systems, reduced the drag forces which today can be found in almost every conceiv- and the cost of locomotion, and increased the effi- able aquatic habitat, ranging from hot soda springs ciency of respiratory and cardiovascular adapta- where temperatures may exceed 40°C to the polar tions during swimming. regions under the ice sheet with temperatures 3 The most important characteristic of water is below 0°C. Despite salt water covering about 70% that it is an almost universal solvent, containing of the Earth’s surface and comprising about 97% complex mixtures of salts, organic compounds and of all water (Horn 1972), as many as 10000 fish gases. Many of these are essential for life and are species (nearly 40%) live in fresh water. The largest taken up by fish using specialized organs in com- numbers of these are found in the tropical regions, plicated processes of ion regulation, osmoregula- 72 Chapter 4 tion, gas exchange and acid–base balance, or via di- mackerel (Scomber scombrus) which has a body gestion of food. just slightly heavier than sea water (1.02–1.06g 4 The transparency of water is very poor. In cm-3), while the heavier bonito (Sarda sarda), clear water, light can penetrate to a maximum skipjack tuna (Katsuwonus pelamis) and bullet of 1000m; however, penetration is usually much mackerel (Auxis rochei) (body densities 1.08–1.09 less than this. Since most food production is gcm-3) have swimming speeds of about 2BLs-1. within the photic zone, most fish reside in this None of these species have a functional swim- region, where prey capture depends on good bladder (Jobling 1998). Continuous effort to sup- visibility and well-developed eyes. In the dark port the body by muscular effort alone would be zone of the oceans, however, fish communicate energetically costly (Marshall 1966). Therefore, and capture prey by producing their own light most other fish, except benthic species that rest on using photophores. the bottom and only swim occasionally, use static In the following sections the physiological lift as a solution to the problem of buoyancy. mechanisms that fish use to cope with these four physical and chemical constraints upon life are 4.2.3 Static lift dealt with in more detail. Fish reduce their specific gravity in a number of ways. 4.2 BUOYANCY, OR 1 The ratio of heavy to light tissue is reduced by COPING WITH PRESSURE minimizing elements of the skeleton, which has a specific gravity of about 3gcm-3. Compared with 4.2.1 Defining the problem terrestrial animals the skeletal elements of fish are Most fish have body densities slightly higher than very much reduced. For example, the skull of fish the specific gravity of sea water (1.026gcm-3). may contain up to 30% lipids, whereas that of Without any lift mechanism they will sink to the mammals rarely contains more than 1% lipid by bottom. Neutral buoyancy allows fish to mini- weight. 2+ 2- mize the energetic cost of staying at a particular 2 Replacement of heavy ions (Mg , SO4 ) with + - + depth, and thus they are able to allocate more light ions (H , Cl , NH4 ) is not a very efficient energy for other activities such as feeding, growth, solution for adult fish. However, it occurs during reproduction, hiding and migration. early development in pelagic eggs (Fyhn 1993; Fish achieve neutral buoyancy and float by Terjesen et al. 1998). Elasmobranchs employ urea using one or both of two physical principles: dy- and trimethylamine oxide (TMAO) as organic namic lift and static lift. osmolytes to reduce their density. During the early larval stages of pelagic marine bony fish, urea is 4.2.2 Dynamic lift produced via both the urea cycle and uricolysis (Chadwick and Wright 1999; Terjesen et al. 2000). Dynamic lift is widely used by elasmobranchs and Although present at considerably lower concen- active teleosts such as mackerels and tunas, which trations than in elasmobranchs, this urea could have bodies that are heavier than water. Lift is gen- possibly lower the density of the larvae, which erated by the caudal and pectoral fins acting as lift- need to reach the upper photic zone where feeding ing foils at about the centre of gravity of the fish begins. and depends on a minimum cruising speed to pre- 3 A better solution to the problem is the removal vent the fish sinking. In scombrid and thunnid fish of ions from the body fluids to the medium, a species a direct correlation exists between the den- process called hypo-osmotic regulation. In the ab- sity of body tissues and sustained long-term swim- sence of gills, kidneys and gut, which are hypo- ming speed. Swimming speeds of about 1 body osmotic regulatory organs found only in the adult, length (BL)s-1 have been recorded for the Atlantic ionic regulation requires the presence of other ap- Physiology of Living in Water 73 propriate structures. In pelagic eggs, fish larvae cholesterol. The liver of these fish contains up to take up water into the rudimentary gut over the 80% squalene and accounts for up to 25% of body opercular pores just before hatching (Mangor- weight. Eggs of these sharks also contain squalene. Jensen and Adoff 1987). Some sharks use sub-neutral buoyancy: they Fish obtain static lift by the use of two different maintain a constant lipid content, which provides materials: lipids and gas. Since gases are much some lift but not enough; they therefore have to more efficient than lipids with respect to providing swim to keep buoyant. lift, fish with swimbladders have no problems in supporting their heavy body components. Fish Gas using lipids for static lift have to reduce their spe- cific gravity as outlined above. Gas is the most efficient material for providing lift, and most teleosts possess gas-filled swimbladders. Swimbladders have sizes that provide neutral Lipids buoyancy: about 5% of body volume for marine The main advantage of using lipids as a means of fish and about 7% of body volume for freshwater static lift is that the lift provided varies very little fish. Since the volume of gas is inversely propor- with depth. Thus if a fish is neutrally buoyant at tional to the ambient pressure, which increases by the surface of the sea, it will also remain close to 101kPa (1atm) for each 10m of depth (Boyle’s law), neutral buoyancy at considerable depths. This the swimbladder must have mechanisms for effi- allows the fish to perform fast vertical migrations. cient filling and emptying at different depths. The main drawback in the long term is that fat and Volume changes are much larger near the surface oils may also be used as fuel for sustained swim- than at greater depths. A fish at the surface ming or even as substrates for growth and develop- (101kPa) halves the volume of its swimbladder ment. Where lipid is the only source of static lift, at 10m depth (202kPa), while a fish at 200m fish will also have great problems adjusting buoy- (2020kPa) has to descend to 400m (4040kPa) to ancy in response to short-term density changes halve the volume. Buoyancy control near the due to feeding and parturition. The most impor- surface will thus be very difficult using a swim- tant lipids are: bladder, and the complete loss of the swimbladder 1 acylglycerol (fatty acid + glycerol, density or at least a reduction in its importance is often 0.90–0.93gcm-3); seen among surface-dwelling pelagic fish. A simi- 2 wax esters (long-chain fatty acids + long-chain lar loss of the swimbladder is found among many fatty alcohols, density ~0.87gcm-3); bottom-dwelling species, where buoyancy control 3 squalene (density ~0.86gcm-3). is of no importance. In general, lipid stores may contribute relatively The swimbladder arises during ontogeny from little to buoyancy in the majority of bony fish a diverticulum in the roof of the foregut. In species. Triglycerides present in the muscles or the physostomatous (Greek physa, bladder; mesenteric fat as a consequence of metabolic lipid stoma, mouth) teleosts the connection between storage may provide some lift for scombrids and the foregut and the swimbladder is maintained, as herring (Clupea harengus). Wax esters in muscles, in elopomorphs and clupeoids, and gas can enter or swimbladder and bones may play some role in be released via this duct. In the more advanced many families of mesopelagic teleosts. Squalene, physoclistuous (Greek kleistos, closed) teleosts which is an extremely light unsaturated hydrocar- this connection has been lost or, for some species, bon, is very important for providing lift, especially closed during development after allowing the in elasmobranchs and particularly pelagic sharks, larvae to fill the swimbladder for the first time. which have large livers and small pectoral fins. Three major problems need to be solved Squalene is formed by the condensation of iso- before the swimbladder can be used for buoyancy prene units in the metabolic pathway leading to regulation: (i) release of gas from the swimbladder; 74 Chapter 4

(ii) maintaining gas inside the swimbladder; (<15m) to nearly 400mgcm-2 in deep water (iii) filling the swimbladder with gas. Gas can be (>1000m, pressure >10100kPa). The oval chamber released by two mechanisms: via the pneumatic can be closed off by nervous control, and gas re- duct into the gut whilst ascending/diving (physo- lease via the circulatory system is regulated in an stomes), or into the circulation via oval chamber advanced countercurrent system called the rete control by nervous stimulation (physostomes/ mirabile (Fig. 4.1). The rete consists of a tight bun- physoclists). dle of afferent and efferent capillaries surrounding The walls of the swimbladder are impermeable each other. This system may be diffuse, such as the to gas due to a thin lining of overlapping guanine micro-rete in salmonids, or it may be at some dis- crystals and other purines. The density of purines tance from the gas gland as in physostomes. In varies from about 20mgcm-2 in shallow waters physoclists the length of the rete is directly corre-

pH = 7.6 pH = 7.8 pO = 5.9 kPa pO = 5.3 kPa 2 2 Oval patch Glucose Glucose

Hb 1 µm PPS

Lactate Lactate TCA

O2 O2 Lactate, H+ H+ H+

– – + HCO3 HCO 3 HCO 3 + H CO2, Triose Swim Guanine crystals bladder

CO2 CO2 + CO2 CO2 H2O

ROOT ON O O HbO2 Hb + 2 2

N Lactate: salting out 2 N2 Gas gland pH = 7.1 Rete pH = 7.3 pO2 = 40 kPa pO2 = 37 kPa PPS, pentose phosphate shunt TCA, tricarboxylic acid cycle

Fig. 4.1 Summary of the main processes involved in filling the swimbladder with gas. The blood entering the descending rete is acidified with lactate, H+ and carbon dioxide from the gas gland. This causes the release of oxygen from haemoglobin via the Root effect (see text). Oxygen and nitrogen, which have been salted out, diffuse into the swimbladder. Multiplication of the solutes in the descending rete is obtained by diffusion from the ascending rete. Oxygen is retained in the rete because the process of binding oxygen is slower than the release of oxygen from haemoglobin (Hb). Physiology of Living in Water 75 lated with the depth at which the fish live and the nurse sharks and Atlantic gadoids when swim- gas pressure in the swimbladder. ming slowly (Wardle and Reid 1977). Steen (1971) solved the puzzle of how the swim- 2 Subcarangiform: species that undulate their bladder of the eel (Anguilla anguilla) is filled with bodies into less than one full wavelength, yet more gas against huge pressure gradients. The gas-filling than one-half body length at speeds greater than mechanism is linked to the countercurrent struc- 1BLs-1. ture of the rete, which allows gas pressure, and 3 Carangiform: species that undulate their bodies therefore volume, to ‘multiply’ within the swim- into a shallow wave up to one-half wavelength bladder (Fig. 4.1). Lactate and carbon dioxide se- within the body length, with the amplitude creted from the gas gland near the luminal end of increasing from head to tail. Typical for fast the rete decrease blood pH, which results in a fast swimmers such as jacks, drums snappers, tunas unloading of oxygen from a very pH-sensitive and mackerels. haemoglobin component in the blood and thus an 4 Ostraciform: these fish swim mainly by flexing increase in the oxygen partial pressure. This so- the caudal peduncle. Typical for fish that need called Root-off mechanism is accompanied by a armour for protection like boxfishes, trunkfishes salting-out mechanism, which allows more and cowfishes. oxygen, but also other gases like nitrogen, to accu- 5 Swimming with fins alone: these fish swim by mulate. In this way oxygen, and also nitrogen, dif- moving their ray-and-membrane fins individually fuse into the swimbladder. Since lactate and rather than their bodies. An example in the lion- carbon dioxide diffuse from the efferent capillaries fish (Pterois volitans). to the afferent capillaries by following a downhill In the following I concentrate on some of the gradient like oxygen, pH again increases in the features of the motors, the drag forces and the cost efferent capillary leaving the rete. Oxygen now of speed, and efficiency of respiratory and cardio- binds again to the haemoglobin component and vascular adaptations during swimming. the oxygen partial pressure decreases. However, this process, which is referred to as the Root-on 4.3.1 Motors mechanism, is much slower than the root-off mechanism, allowing a fraction of the oxygen Rhythmic undulations of the body or fins exert content to remain within the system. force on the relatively incompressible surrounding water and the fish will move either forward or backward. The undulations result from the con- 4.3 SWIMMING traction of the segmented axial musculature, which is divided into myotomes by the myoseptal– Though some fish can fly, crawl on land, climb connective tissue partitions on which the muscle trees or burrow in mud, most fish swim primarily fibres insert. Since the vertebral column is by oscillating their bodies. Some swim by moving incompressible, these alternating contractions in just their paired and unpaired fins (‘propellers’). the myotomes on each side result in lateral bend- The analysis of swimming is very complex, and ing of the body. The muscle fibres of the myotomes certainly needs more space than can be given in lie in a series of horizontal tubes of connective tis- this basic approach. There is of course great varia- sue, which run along the length of the fish. Most tion in how fish swim due to the large variability of of the myotomes lie ventral to the vertebral col- their body shapes. These swimming methods can umn and form a characteristic pattern, which in be divided into five basic types. amphioxus and fish larvae looks like the letter ‘V’ 1 Anguilliform: whole body is flexed into lateral pointing towards the tail. In all other adult fish the waves for propulsion. Typical for ‘eel-like’ species overlapping myotomes are folded into ‘W’ shapes. such as eel, marine gunnels and lampreys. Other In higher fishes, only the fibres just under the skin species that use this type of swimming include lie along the axis, while the deeper fibres in the 76 Chapter 4 myotomes spiral up to 30° to the axis. This bluefish, striped bass and herring, while others, arrangement assures a near-isometric contraction such as the chub mackerel, rainbow trout, coalfish at the lowest cost, since the deeper fibres all con- and carp, show a broader range of white muscle tract to a similar degree for a given bending of the activity (Freadman 1979). trunk, so the longitudinally running fibres shorten about 10% during swimming while the deeper fi- 4.3.2 Drag forces and the cost bres only shorten about 2%. The more undulatory waves the fish can exert against the surrounding of speed water, and the faster and more exaggerated the Swimming is retarded by two kinds of drag forces: waves are, the more power can be generated, re- 1 pressure drag, which is caused by the vorticity in sulting in higher accelerations and swimming the wave creating a higher pressure at the nose speeds, given that the resistance is held constant. than at the tail and which is highest for unstream- lined fish; and Slow and fast muscles 2 skin friction drag, which is caused by the viscos- ity of water and which makes it stick as a boundary Just below the skin of the fish there is a thin layer of layer to the fish surface. myoglobin-rich red fibres that covers most of the Skin friction drag is the most important drag com- myotomes, which consist of white fibres. Fish use ponent for fish. For the fish to keep moving, the in- these red fibres, which are commonly referred to as ertial forces of motion must be larger than the slow fibres, for cruising. The main mass of white viscous forces of the surrounding water. The di- fibres is used for burst swimming. The main char- mensionless Reynolds number, Re, describes the acteristics of red and white fibres are described relationship for theses two forces in fish: in Table 4.1 (Bone 1978). Guppy and Hochachka (1978) showed how changes in muscle pH, oxygen VL Re=>()always 1 , (4.1) pressure, temperature and biochemical substrates v can reversibly alter aerobic and anaerobic enzyme activities in skipjack tuna white muscle. Thus where V is main swimming velocity, L the length some fish conform to the classical differentiation of the fish and v the coefficient of kinematic vis- of white and red muscle function, for example cosity of the fluid (v = viscosity/density, and for

Table 4.1 Comparing slow red and fast white muscle fibres in fish myotomes.

Red fibres White fibres

Diameter about 60–150mm Fibres more than 300mm Rich vascularized Poorly vascularized Abundant myoglobin, colour red No myoglobin, colour white Abundant large mitochondria Few smaller mitochondria Oxidative enzyme systems Enzymes of anaerobic glycolysis Low activity of Ca2+-activated myosin ATPase High activity of Ca2+-activated myosin ATPase Little low molecular weight Ca2+-binding protein Rich in low molecular weight Ca2+-binding protein Lipid and glycogen stores Glycogen store, usually little lipid Sarcotubular system has lower volume than in fast fibres Relative larger sarcotubular systems Distributed cholinergic innervation Focal or distributed cholinergic innervation No propagated muscle action potentials Propagated action potentials; may not always occur in multiply innervated fibres Long-lasting contractions evoked by depolarizing agents Brief contractions evoked by depolarizing agents Physiology of Living in Water 77 water ª0.001 and is constant at any given tempera- Fuelled by anaerobic and aerobic metabolism. The ture). For fish swimming normally, Re is within so-called critical speed is calculated from the the range 104–108. In this range viscosity effects are maximum speed achieved prior to fatigue in a mostly confined to a thin boundary layer adjacent tunnel respirometer, where the speed has been to the body surface. At Re < 5 ¥ 105, viscous forces increased stepwise until fatigue is reached. dominate and the boundary layer on a flat plate 3 Burst swimming: burst speeds that can be main- will be laminar. At higher Reynolds numbers tained for less than 20s. Characterized by an initial the boundary layer will be turbulent. Thus fast- acceleration phase followed by a sprint phase of swimming fish have to cope with a larger skin steady swimming defined as swimming in a given friction drag. This drag can be lowered by reducing direction at constant speed. This type is predomi- the wetted area and lateral movements, but most nantly fuelled by anaerobic metabolism. importantly by a boundary layer control mecha- Tunnel respirometry of fish calculates the cost nism. These mechanisms mainly tend to delay the of transport for swimming at a given speed in transition from a laminar layer to a turbulent layer Jg-1 km-1 as well as its most efficient speed (Brett so keeping Re low. A mucous coating acts in this 1975). The locomotion of fish is very efficient com- way. It has also been suggested that opercular pared with that of animals on land. Transport in water flow may smooth the flow of boundary water is thus about ten times less costly than water next to the body. Sharks have denticles with transport in terrestrial animals (Schmidt-Nielsen low sharp-edged ridges parallel to the swimming 1997). However, with increased swimming acti- direction that are believed to delay transition by re- vity, tissue demand may increase 5–15-fold, with ducing microturbulence in a laminar boundary about 90% of this increase being accounted for layer. In addition many fast-swimming fish, such by elevated muscle metabolism. It is reason- as scombrids and the blue shark, can save up to able to believe that the efficiency of transferring 20% of the energy needed for locomotion by using momentum from the body and fins to the water in a so-called glide–power cycle, where ‘propulsive’ scombrids may be as high as 85% (Bone et al. 1996). body movements are alternated with periods of Webb (1975) showed that rainbow trout expend gliding when the body is held rigid. 18% of the total work of acceleration to overcome frictional drag. A number of respiratory and car- diovascular responses are invoked in order to 4.3.3 Efficiency of respiratory meet this demand for increased oxygen supply. In- and cardiovascular adaptations creased swimming rate results in a minor increase during swimming in breathing rate and a significant increase in ven- tilatory stroke volume. As a result ventilation Traditionally the different taxonomic groups have volume may increase eightfold, and the contact been divided into ‘cruisers’ that mainly use red time between water and respiratory exchange muscles, and ‘burst swimmers’. However, in order surface decreases from about 300ms at rest to to discuss efficiency of respiratory and cardiovas- about 30ms when active. At the same time, cular adaptations during swimming we have to be cardiac output is increased mainly by a 50–250% more precise about defining the term ‘swimming’. increase in stroke volume and a 20–40% increase In fish physiology three main categories of swim- in heart beat frequency. Though the blood transi- ming are defined (Jobling 1998). tion time in the gill lamellae is reduced from 3s to 1 Sustained swimming: speeds that can be main- 1s, gill recruitment is increased due to higher tained for 200min or longer. Totally aerobic blood pressure. As a result of these respiratory and metabolism. cardiovascular adjustments more oxygen is chan- 2 Prolonged swimming: speeds that can be main- nelled to the muscles to meet the increased tissue tained for 20s to 200min and result in fatigue. demand. 78 Chapter 4

pensated for by salt intake in the food and by a 4.4 OSMOREGULATORY high-affinity salt-uptake mechanism at the gills, PROBLEMS IN FRESH AND while water balance is maintained by the excre- SALT WATER tion of about 2–6mlkg-1 h-1 of copious urine (10– 4.4.1 Defining the problem 80mosmol). Drinking rate is kept low. Fish live in many different environments, ranging Processes in the gills from virtually distilled water to salt concentra- tions that are so high that living underwater may The gills are the site of many of the exchange be difficult. Though most fishes are stenohaline, processes that occur between the fish and its envi- meaning that they live all their life in water of ronment, such as gas exchange, ion regulation in nearly constant salt concentration, many fish are the form of active uptake of sodium and chloride able to move between fresh water and sea water. ions, acid–base regulation, and excretion of Such fish are euryhaline and/or diadromous. nitrogenous waste products of protein catabolism Catadromous species, such as eels, leave fresh such as ammonia and ammonium ions. Since water to spawn in the sea, while anadromous these various exchange mechanisms are coupled, species, such as salmon, which are more numer- the excretion of waste products may result in the ous, make the reverse migration to spawn in fresh uptake of sodium and chloride (Fig. 4.2). Thus chlo- water (see Metcalfe et al., Chapter 8, this volume). ride is taken up in exchange for bicarbonate, and The skin of fish is relatively impermeable to water sodium is taken up in exchange for protons and/or and ions. The gills, however, have large areas of ammonium ions. In this respect the double permeable epithelium, which are in contact with exchange mechanisms seem to differ from those the surrounding water. The osmolality (number of present in marine teleosts (i.e. Na+/K+/2Cl- undissociated solute molecules or ions per kilo- cotransport, see Section 4.4.3). Thus, the double gram of solvent) of fresh water is <10mosmol, and exchange mechanism provides for: sea water falls within the range 800–1300mosmol. 1 maintenance of appropriate internal [Na+]; The freshwater teleosts have blood osmolalities of 2 maintenance of appropriate internal [Cl-]; -1 + about 260–330mosmolkg . Thus these fish will 3 elimination of toxic NH3 ; - be subjected to a passive influx of water and efflux 4 elimination of CO2 as HCO3 ; of ions. In marine fishes, which have blood osmo- 5 adjustment of internal H+/OH-; lalities in the range 370–480mosmolkg-1, these 6 maintenance of internal ionic electrical fluxes will be reversed. In addition to osmoregula- balance. tion, many fish regulate the ionic composition in their blood by active, energy-consuming pro- Processes in the kidneys cesses. In the following sections I examine in more detail how different groups of fish regulate their The elongate teleost kidneys are, like adult lam- blood osmolality and ionic composition to the sur- prey kidneys, composed of typically segmented rounding water. mesonephrons (Fig. 4.3). They consist of glomeruli in the posterior lobe of the kidney, which receives 4.4.2 How to regulate in fresh water oxygenated blood from the dorsal aorta, and proxi- mal and distal tubules in the more anterior part of Since all freshwater fish have blood osmolalities the kidney that receive venous blood from the higher than the surrounding water, they have caudal vein. evolved mechanisms to (i) prevent the passive The glomerular filtration rate (GFR) is high in efflux of ions through diffusion over the body and freshwater teleosts. They produce about ten times loss from urine and faeces and (ii) remove excess more urine than marine teleosts. The urine is water. In lampreys and teleosts, ion loss is com- much more dilute than the plasma and as much as Physiology of Living in Water 79

B + H+ HB pH =~ 8 – + H+ CO w HCO3 2 ~ – + pHw = 7.2–7.7 [HCO3]2.5 mM NH3 + H + – NH4 [HCO3]2.5 mM Cl – Na+

Water . – + + –1 –1 HCO3 NH3 + H NH4 VG = 12000ml kg h α–AA Gill CA +H2O O HCO – CO H+ . 2 3 2 Q = 10–100ml kg–1 h–1

RBC O H OCl– HCO – Na+ CO Carbamino 2 2 3 2 swelled HHb HbO2 K+ CO CO 2 H+ Hb– 2 O H CO – – 2 CO2 + H2O 2 3 HCO3 Cl H2O

Dissolved CA H OCl– HCO – 2 3 H2CO3 Dissolved CO2 + H2O – O2 Hb +H CO 2 CO 2 K+ RBC HbO2 HHb Carbamino Blood CO + – – 2 Na HCO3 Cl H2O O2 Tissue CO2 O2

Fig. 4.2 Summary of the main ion transport and acid–base regulatory processes taking place in the gills in fresh water and its relation to the transport of oxygen to, and carbon dioxide from, the tissues. At the top, a secondary lamella is in contact with the water, which flows from left to right. Beneath is a blood vessel with red blood cells

(RBC). The left cell illustrates the processes in the gills (loading O2, releasing CO2); the right cell shows the processes in the tissues (delivering O2, loading CO2). The gills are the main site of acid–base regulation mainly due to the high flow of water over the gills, resulting in a high buffer (B) capacity. The uptake of sodium and chloride ions is directly coupled to the excretion of carbon dioxide and ammonia. Most of the carbon dioxide is transported as bicarbonate in the blood, a minor fraction as physical dissolved carbon dioxide and some directly bound to the haemoglobin molecule. Some of the carbon dioxide entering the red blood cells diffuses out into the plasma. Electrical neutrality is maintained by the inward diffusion of chloride ions followed by water, which causes the cell to swell (referred to as the chloride shift). The binding of carbon dioxide and protons to haemoglobin facilitates oxygen release. to the tissues, while the release of carbon dioxide over the gills facilitates the uptake of oxygen by haemoglobin. V , flow of water . G over gills; Q, cardiac output (blood flow); CA, carbonic anhydrase; a-AA, alpha amino acids. (Source: acid–base part based on Heisler 1984.)

99.9% of the sodium and chloride ions passing into the monovalent ions occurs in the distal tubules, the glomerular filtrate are resorbed. The resorp- which are impermeable to water. Urine produc- tion of organic solutes occurs primarily in the tion is proportional to GFR. The urine is collected proximal tubules. This resorption is accompanied in the urinary bladder and urination may occur at by some water uptake. Most of the resorption of intervals of 20–30min. This structural arrange- 80 Chapter 4

Glomerulus Proximal tubule Distal tubule Water Monovalent ions Glucose, macromolecules, amino acids Divalent ions Monovalent ions

Impermeable to water

– + GFR=10mlkg–1h–1 HCO3 + H CO2 +H2O 20–80mOsmol 220–320mOsmol B + H+ HB ~ pH <6 pHu=7.8 u – + + [HCO –] <0.1mM [HCO3 ]u>5mM Na+ H NH3 NH4 3 u

+ + α NH4 H + NH3 -AA Bicarbonate + + HCO – conservation H 3 CO2 Blood supply Glomerulus: Dorsal aorta Tubules: Caudal vein

Fig. 4.3 Summary of the main functions of the kidney in freshwater fish, showing the major sites of absorption and water permeability. The kidney does not play the same role as the gills in acid–base regulation due to the considerably lower flow and thus buffer capacity. GFR, glomerular filtration rate; B, buffer; a-AA, alpha amino acids. (Source: based on Lahlou 1980; Heisler 1984.) ment of the kidneys in freshwater teleosts allows 2 osmolality close to sea water, internal inorganic considerable control of the composition of the ion concentration about one-third of sea water glomerular filtrate, but little control over the final (elasmobranchs, Latimeria); urine volume. The freshwater stingrays from the 3 osmolality of about 400mosmol, internal inor- Amazon (Potamotrygon sp.), which live more than ganic ion concentration about one-third of sea 4500km from the sea, also have a very high GFR. water (marine teleosts). Urea is reduced to only 1mm, and sodium and chloride ions are reduced to similar levels (Bone Hagfish et al. 1996). Hagfish have a blood volume twice that of 4.4.3 How to regulate in sea water Gnathostomata. The blood is isosmotic to sea water and concentrations of sodium and chloride Sea water has an osmolality of about 1000mosmol. ions are the same as in sea water, but are slightly Among marine fish three different strategies for higher in the tissues due to high levels of intracel- regulation of internal water and total solute con- lular amino acids. Osmotic water exchange is centrations can be distinguished: more or less absent, while water exchange rates 1 same osmolality as sea water, no regulation at may be as high as 2.3lkg-1 h-1 (McInerney 1974) all (hagfish); due to the free permeability to water. Urine pro- Physiology of Living in Water 81 duction is minimal, and chloride cells have been produce urine more concentrated than the blood so observed. Though hagfish are strictly marine that excess salts are removed and water is con- (stenohaline), experiments by McInerney (1974) served. In fact many marine teleosts have evolved demonstrated that they have the ability to reab- an aglomerular kidney to minimize water loss, and sorb sodium from the glomerular filtrate, a feature urine flow rate is as low as 1–2% of body weight per necessary for living in fresh water. This could be day. Marine fish have solved the osmotic problem reminiscent of an original freshwater ancestry as follows (Potts 1976; Kirsch et al. 1981). After (Marshall and Smith 1930). drinking, the water passes into the oesophagus, which is impermeable to water but permeable to light monovalent ions. Sodium and chloride thus Marine elasmobranchs diffuse into the blood down their concentration Like the hagfish the elasmobranchs are not in gradients. When entering the intestine, which is danger of losing water because they are isosmotic permeable to water, the water has become hypos- to the medium. However, like the marine teleosts motic to sea water and nearly isosmotic to the they are hypo-ionic to sea water. Plasma os- blood. A Na+/K+/2Cl- cotransport mechanism molarity is increased to seawater levels by the located in the apical brush border membrane trans- organic osmolytes, urea and TMAO. Urea is pro- ports the ions from the gut lumen into absorptive duced from the nitrogen waste product, ammonia, cells and water follows passively in a process called in the ornithine–urea cycle in the liver, from where solute-linked water transport. Na+/K+ ATPase, it enters the plasma and other body fluids. Though which is present in the basal membrane of the urea is less toxic than ammonia, the levels seen in absorptive cells, provides the energy for this active elasmobranch plasma would be fatal for other fish transport. The excess sodium, potassium and chlo- except Latimeria. Elasmobranchs have achieved ride ions are then transported in the blood to the higher tolerance to urea by retaining TMAO, gills where they are excreted via the ‘chloride cells’ which counteracts the effect of urea at a ratio of (Fig. 4.4). The concentration of divalent ions is low about 1:2 in their body fluids, and by having en- in sea water; 80% of these ions that are taken up by zymes and proteins that appear to be less sensitive drinking are expelled in the faeces. The remaining to disruption by urea. A special sodium chloride ions are excreted in the urine. excretory organ has evolved in the elasmobranchs. In order to become truly marine the teleosts had This so-called rectal gland uses a Na+/K+/2Cl- co- to solve the osmotic problems for the spawned transport mechanism to secrete excess sodium egg. At spawning, marine fish eggs must contain a and chloride ion intake from ingested food. The water reservoir to compensate for the passive secretion contains very little of the organic water loss imposed by the hyperosmotic sea water osmolytes and is isosmotic to the plasma. Energy (Fyhn et al. 1999). The high water content of the for the cotransporter is provided by an abundance yolk of marine teleost eggs forms this water of Na+/K+ ATPase in the gland. reservoir. Regardless of systematic affinities, most extant marine fishes spawn pelagic eggs. Yolk protein hydrolysis and increase in content of free Marine teleosts amino acids during final oocyte maturation is part In contrast to marine elasmobranchs the marine of the mechanism that brings water into the yolk teleosts, being hyposmotic to sea water, lose water before the eggs are spawned (Rønnestad et al. 1996; to the medium primarily over the thin gill epi- Thorsen & Fyhn 1996). thelium. They have to replace this water loss and drink between 10–20% and 35–40% of their body 4.4.4 Antifreeze weight per day. In addition there is a constant pas- sive influx of monovalent ions over the gills as well For hagfish, marine elasmobranchs and freshwater as dietary salt uptake. The kidneys are not able to teleosts, freezing is not a problem since they have 82 Chapter 4

– Internal Cl 150 mM Na+ Secondary lamellae ~ Negative charge 3 mM K + ~ – ~ 140 mM Cl

Na+ K + – Sea water Na+ K + 2Cl 470 mM Na+ ~ + Blood ~ 10 mM K channels 560 mM Cl– ~ Positive Site of α-chloride cell: charge Na+-K+-ATPase in fresh and sea water Apical pit Fig. 4.4 Schematic illustration of Pillar Tight junction part of secondary lamellae on a gill cells Leaky junction filament showing the two types of chloride cells involved in the active secretion of chloride ions. Accessory The process depends on Na+/K+ cell Gill filament ATPase activity and causes β-chloride cell: polarization of the cell. Sodium in fresh water ions may diffuse out via the leaky junctions.

body fluids that are either isosmotic or hyperos- motic to the surrounding water, as long as the 4.5 RESPIRATION AND water is not frozen. Marine teleosts, however, are SPECIAL ADAPTATIONS FOR hyposmotic to the ambient water and can freeze to LIVING IN LOW OXYGEN death when the water temperature drops below 4.5.1 The environment 0°C. To prevent freezing the fish may produce macromolecular ‘antifreeze’ compounds in their Fish live in a dense and very viscous environment, blood serum. The antifreeze consists of glycopro- which for a terrestrial animal would appear to be teins or proteins, which coat and isolate ice crys- almost depleted of oxygen. At best, one volume of tals forming in the blood by binding their hydroxyl water contains a little less than 4% the amount of groups to oxygen molecules on the surface of the oxygen present in the same volume of air. In air ice crystals (DeVries and Wohlschlag 1969). The the solubility of a gas would be roughly 1000ml/ production of antifreeze is dependent on cold accli- 101.3kPa = 9.87mll-1 kPa-1 (at 15°C). In practice mation and short photoperiods (Duman and De- the respiratory gases are treated as ideal gases, Vries 1974a), and genetically based differences which means that oxygen and carbon dioxide in antifreeze production have been described would dissolve more or less equally in air. In water, (Duman and DeVries 1974b). In the aglomerular however, the solubility of gases depends on: (i) the kidneys of Antarctic fish the glycoproteins are nature of the gas, (ii) the pressure of the gas in the conserved rather than filtered out of the blood, gas phase, (iii) the temperature and (iv) the pres- which lowers the energetic cost of osmoregulation ence of other solutes. Thus, while the solubility of in these fish (Dobbs et al. 1974; see also Section oxygen in water is only about one-thirtieth the sol- 4.5.4). ubility in air, the solubility of carbon dioxide is Physiology of Living in Water 83 more or less the same in air and water depending on 4.5.2 Gas exchange over gills temperature and is similar at about 15°C. For calculating gas fluxes we use the concept of All fish have a unidirectional flow of water over Ohm’s law: flux = capacitance ¥ potential differ- their gills, except for adult lampreys, which have ence, which gives the following three fundamental tidal ventilation in and out of the gill sac via the gas equations: valved branchial openings. The mouth cannot be open since the lamprey is likely to be attached to ˙˙ either rocks or hosts. Synchronous expansion and MVCO2 =◊-mei[]b m () PPCO (4.2) 2 contraction of the buccal opercular cavities venti- ˙˙ late the gills. This provides a nearly constant flow MVOm2 =◊-mie[]b () PPO (4.3) 2 of water over the gill surfaces. At a given swim- . ming speed fish adopt ram ventilation to conserve where Mgas is the volume of gas taken up (O2) or . energy. The branchial pump is switched off and the produced/released (CO2), Vm the flow of medium flow of water over the gills is regulated by the open- over the gills, bm the capacitance coefficient (solu- bility) of the gas in the medium and P the partial ing of the mouth in relation to the actual speed and pressure of gas (i, before gills; e, after gills). From oxygen demand. The gills consist of bony or carti- these expressions, we may calculate: laginous arches to which one row of paired gill filaments are anchored. Numerous secondary ˙ lamellae protrude from both sides of each fila- MCO 2 = RQ (4.4) ment. A layer of thin epithelial cells covers the M˙ O2 outside of the lamellae. Beneath the basement membrane are supportive pillar cells and blood where RQ is the respiratory quotient (R, the ex- vessels running in the opposite direction to the change ratio, is used when examining isolated water flow (Fig. 4.4). This arrangement ensures parts of the fish). Rearranging equations 4.2–4.4 very efficient gas exchange, with oxygen utiliza- and ignoring P : i tion as high as 80% and postbranchial oxygen par- tial pressures higher than in the water passing ()bm O2 through the gill. The functional area of the gills can ()PPPe = ◊-()ei (4.5) CO2 O2 ()bm be changed by shunts in the secondary lamellae or CO2 by the action of an autoregulatory system, which

If the fish extracts all the oxygen in the water (Pi @ reacts on blood pressure by increasing the tonus 21kPa, Pe = 0kPa), the partial pressure of carbon of the contractile elements in the pillar cells so dioxide can, at a maximum, be about 0.7kPa in the re-routing blood flow (Laurent 1984). Intrinsic exhalant water. This means the maximum partial muscles in the gill filaments of most teleosts can pressure of carbon dioxide in the arterial or post- change the angle of the filaments on each arch and branchial blood will not exceed this value, which thus alter the flow pattern over the secondary is almost a factor of 10 lower than for air-breathing lamellae. animals. This low oxygen availability has undoubtedly Other sites for gas exchange contributed to the evolutionary development of: 1 large gill surface areas for extremely efficient gas An efficient gas exchange system should have exchange; the following characteristics: close contact to the 2 air-breathing organs and the necessary circula- medium, short diffusion distances, and a well- tory arrangements; developed circulation that can transport oxygen 3 acid–base regulation depending more on ion efficiently from the site of gas exchange to the tis- than ventilatory exchanges. sues. Cutaneous respiration may be of some signif- icance in a few cases either when oxygen demand 84 Chapter 4 is very low, as in some Antarctic fishes, or when or other changes. Aneural control is mainly ef- the diffusion distance is very short, as for fish fected by: larvae. In adult fish, cutaneous respiration is 1 changes in blood volume and venous return by less than 30% of routine metabolism (O. Brix, swimming movements and mobilization of blood unpublished results). Its significance is still poorly into general circulation from the spleen, liver or investigated, but it may be of some importance for blood sinuses in various species; fish like eels when migrating short distances 2 direct response of heart muscle (pacemaker) to across land (Berg and Steen 1965). For air-breathing temperature changes, which increase the heart organs see Section Air breathing. rate (Randal 1970); 3 circulating catecholamines like adrenaline, 4.5.3 Circulation and gas transport which stimulates heart rate and stroke volume and dilates the gill vasculature (Nakano and The main tasks of the circulatory system are to Tomlinson 1967; Bennion 1968), and noradrena- transport respiratory gases, nutrients and metabol- line, which constricts systemic vascular beds ic waste products, endocrine factors and heat. (Wood and Shelton 1975). Most fish have a single circulatory system. The The hearts of teleosts and elasmobranchs, but heart consists of the sinus venosus, atrium, ven- not hagfish and lungfish are innervated by the tricle and bulbus. In elasmobranchs, agnatha and vagus nerve (Xth cranial nerve). Stimulation of holosteans the bulbus is replaced by the conus the vagus causes a cholinergic effect, simulating arteriosus. The conus does not increase the accel- the effects of acetylcholine, which slows the heart eration of blood, as does the bulbus in teleosts. The rate in teleosts and elasmobranchs producing blood is pumped into the ventral aorta, where bradycardia and increases the heart rate in lamprey blood pressure is in the range 4.0–7.3kPa. After (Randall 1970). Light flashes, sudden movements passing the gills, which create ~30% of the total of objects or shadows, and touch or mechanical vi- resistance to blood flow, the blood pressure has brations may increase vagal tone. The response decreased to about 3.3–4.7kPa in the dorsal can be blocked by injection of atropine. Some fish aorta, which flows alongside the caudal vein in also have adrenergic or stimulatory fibres from the the haemal canal formed by fusion of vertebral vagus. processes. The haemal canal thus protects the The circulatory system may be linked to maxi- caudal vein from the pressure waves that move to mization of oxygen uptake in the gills and delivery the caudal end of the fish as the myotomes con- to the tissues. In the following I examine in more tract during swimming. The pressure for venous detail some of the key factors in this process. return is generated in the segmental veins in the muscles in the caudal region by the so-called The significance of haemoglobins haemal arc pump, or in the caudal sinuses or caudal pump, which can be called caudal hearts, Fish haemoglobins are mostly tetramers consist- before entering the caudal vein. Back-flow is ing of two a and two b polypeptide chains (141 and prevented by a series of valves. The caudal vein 146 amino acid residues, respectively). The basic collects blood from the cutaneous circulation, function of the haemoglobins is to ensure an ade- the posterior muscles and the remainder of the quate supply of oxygen to all parts of the organism caudal region (Satchell 1965, 1971). Hagfishes have in which they occur. In order to accomplish this no less than five accessory hearts. For circulatory task they have developed, in the course of evolu- adaptations in air-breathing fish see Section Air tion, a common molecular mechanism based on breathing. the principle of ligand-linked conformational Aneural and neural mechanisms control the change in a multi-subunit structure (Wyman heart and vascular system allowing for a wide 1964). Within the framework of this common range of circulatory adjustments to environmental mechanism, however, different haemoglobins Physiology of Living in Water 85

160 Antarctic fish

80 ) 2 60

40

(ml blood per ml O Tunas

Blood perfusion requirement 20 Fig. 4.5 Blood perfusion require- ments of fish in relation to blood 0 haemoglobin concentrations. 0 5 10 15 (Source: based on Wood and –1 Lenfant 1979.) Haemoglobin concentration (g dl )

have acquired special features to meet special CO2 and nucleotide triphosphates (NTP), called needs. The significance of haemoglobins is illus- ligands (Brunori et al. 1985). Although the product trated in Fig. 4.5, which shows the impact of of the central exon of the haemoglobin gene (the haemoglobin concentration on perfusion require- oxygen-binding site) is sufficient for accomplish- ment expressed as millilitres. of blood pumped by ing the primary function of haemoglobin, the Q the heart (cardiac output, ) for. each millilitre of products of the external exons (the ligand-binding oxygen taken up by the cells (Vo2) for random ex- sites) allow external modulation of the functional amples of fish species. The perfusion requirement, properties (Eaton 1982). Hence, oxygen binding at which links cardiac output and blood oxygen the haem groups is modulated to different extents transport to oxygen uptake, is calculated from the in different species by interactions with ligands flux equations shown in Section 5.1: (Fig. 4.6). Haemoglobins exhibit a great deal of vari- ation, in terms of absolute affinity for oxygen, in their susceptibility to metabolic effectors; these Q˙ Ê 1 ˆ = Á ˜, (4.6) variations allow haemoglobin to fully meet the Vo˙ Ë CC- ¯ 2 aO22 vO physiological requirements of a given species. This type of tuning is primarily based on the ability of where C - C is the difference in oxygen con- effectors to preferentially bind to one of the two aO2 v¯O2 tent between arterial and mixed venous blood. The quaternary conformations. In particular, preferen- Antarctic fish without haemoglobins thus have to tial binding of NTP, typically to the low oxygen pump almost ten times as much blood for the same affinity T-state, facilitates the unloading of oxygen amount of oxygen taken up as the tunas with the to the tissues (Fig. 4.6). The decrease in oxygen highest haemoglobin concentration. affinity caused by protons is commonly referred to as the Bohr effect. When the binding of protons is so strong that the haemoglobin remains in the T- Sites of control state, in which there is no transition to the R-state Oxygen transport is based on conformational and thus no cooperativity, this is called a Root changes of the quaternary structure of the haemo- effect. This plays an important role in filling the globin molecule that are caused by the binding and swimbladder (see Fig. 4.1). Fig. 4.6 summarizes release of small solvent components, such as H+, how the oxygen uptake of fish can be modified by 86 Chapter 4

Blood depot and 30 erythropoietic AV-difference reserve

Ventilatory reserve

) Arterial –1 20

O2 uptake = . AV-difference × Q

Venous content (ml dl 2

10 Blood O Temperature NTP PCO2 pH

0 . . . 0481216 Q0 Q1 Q2 Cardiac output (ml min–1) PO2 (kPa)

Fig. 4.6 The magnitude of oxygen uptake of fish is illustrated as a square, where one side is .the arterial–venous difference in blood oxygen content (AV-difference) and the other side is the cardiac output (Q). The arterial–venous difference can be increased by increasing ventilation and/or the release of more red blood cells into the blood from the spleen (stress response), or by increasing the unloading of oxygen to the tissues under the allosteric control of ligands and temperature (the most important nucleotide triphosphates (NTPs) in fish are GTP and ATP). Cardiac output is the product of stroke volume and heartbeat frequency, and will accordingly be changed in parallel with these parameters.

regulating haemoglobin oxygen transport and car- ity of haemoglobin components appears greatest in diac output. ectothermic animals, and in particular among fishes that may experience both rapid tidal or diur- nal, and marked seasonal, changes in ambient oxy- The significance of haemoglobin multiplicity gen and temperature. It has been suggested that The term ‘haemoglobin multiplicity’ means the both the number of isohaemoglobins and their occurrence of more than one haemoglobin compo- functional heterogeneity is related to the constan- nent in the same or different developmental stages cy of physicochemical conditions in the environ- of a species. Vertebrate red blood cells typically ment (Brix et al. 1999). However, demonstrating contain more than one kind of haemoglobin. this simple hypothesis has been hampered by Commonly, multiple genes are expressed that phenotypic plasticity, which can produce for code for different variants of these globins, result- example an acclimatory response of the individ- ing in ‘isohaemoglobins’. Some species show ual, and by phylogenetic divergence in physiologi- haemoglobin ‘polymorphism’, the occurrence of cal mechanisms. different ‘allohaemoglobins’ in different individ- The selective advantages of heterogeneous uals representing different genetic strains (Brix isohaemoglobins have been considered for very et al. 1998a; Samuelsen et al. 1999). The multiplic- few, predominantly European and North Ameri- Physiology of Living in Water 87 can freshwater teleosts, including fish with atypi- interpretation of experimental results is con- cal diadromous behaviour (Brix et al. 1998b). For founded by phylogeny. example, trout (Salmo trutta) and eel (Anguilla anguilla ) possess both anodic isohaemoglobins 4.5.4 Influence of environmental which are sensitive to temperature and pH and are characterized by strong Bohr and NTP effects, and temperature and hypoxia on oxygen high-affinity cathodal components with low sensi- uptake and transport tivity to pH and temperature (Binotti et al. 1971; Coping with environmental temperature Fago et al. 1995, 1997a). In addition the effects of temperature and hypoxia on the expression of spe- Critical temperature thresholds (Tc) have been cific isohaemoglobins (Murad and Houston 1991; defined as the transition to an anaerobic mode Houston and Gingras-Bedard 1994) demonstrate of mitochondrial metabolism, once temperature the phenotypic plasticity of the haemoglobin reaches low or high extremes, due to insufficient system in trout. Functionally heterogeneous ventilation and/or circulation and aerobic energy isohaemoglobins of many Amazonian fishes have provision (Zielinski and Pörtner 1996; Pörtner been suggested to be adaptive to water oxygen ten- et al. 1998). The lower Tc is set by insufficient aero- sions in the tropics (Powers 1974, 1980; Fyhn et al. bic capacity of the mitochondria, while the upper 1979; Val et al. 1990). This feature is supported by Tc is set by a mismatch of excessive oxygen the observation that haemoglobins from ten dif- demand by mitochondria and insufficient oxygen ferent species of African cichlids share an identical uptake and distribution by ventilation and circula- electrophoretic pattern and similar haemolysate tion (Pörtner et al. 2000a). In the cold, energy de- oxygen-binding properties (Verheyen et al. 1986). mand will be met by mitochondrial proliferation, However, the Antarctic notothenioid fishes have which ultimately will cause an increase in oxygen few isohaemoglobins present, which correlates demand that becomes detrimental at the upper Tc. with low activity levels and the stability of the Oxygen demand is not only related to cellular polar marine environment (high dissolved oxygen energy requirements but also to the number of and constant low temperature) (DiPrisco et al. mitochondria and their properties, specifically 1990; DiPrisco and Tamburrini 1992). Differences maintenance of the proton gradient. The relation- in the functional properties of these simple ship between this so-called proton leakiness and haemoglobin systems is correlated not with envi- aerobic capacity appears to be constant. Pörtner ronmental factors but with organismal factors et al. (2000a) suggest that metabolic cold adapta- such as swimming behaviour and body mass tion may depend upon the extent of diurnal and (Wells and Jokumsen 1982; Fago et al. 1997b). seasonal temperature fluctuations, leading to Haemoglobin is not expressed at all in the chan- higher costs of maintenance in eurythermal than nichthyid fishes (Cocca et al. 1997). Brix et al. in stenothermal fish. Thus in stenothermal fish (1999) showed that triplefin fishes living in aerobic capacity and energy expenditure is mini- shallow, thermally unstable habitats possess a mized as far as possible according to environmen- greater number of cathodal isohaemoglobins, tal and lifestyle requirements. haemoglobin components that migrate towards The haemoglobin polymorphism of Atlantic the negative pole during electrophoresis. These cod (Gadus morhua), which possesses the pheno- species have haemoglobins with higher oxygen types HbI11, HbI12 and HbI22, was described more affinity and reduced cooperativity and which are than 30 years ago (Frydenberg et al. 1965; Sick less sensitive to changes in pH compared with 1965a,b). The frequency of the two alleles HbI(1) haemoglobins of species occurring in more stable, and HbI(2) shows a north–south cline along the deeper water habitats (Fig. 4.7). The analysis of an Norwegian coast: the frequency of the HbI(1) assemblage of closely related species circumvents allele is about 10% in the Barents Sea in Arcto- some of the difficulties inherent in studies where Norwegian cod, 20–50% along the coast of north- 88 Chapter 4

0 10

(m) 20 Depth 30 Cathode – 7.0 6.5 pl 6.0 + 5.5 Anode 5.0 Cathodic haemoglobins 8 Anodic haemoglobins

4 Expressed haemoglobins 0

Oxygen affinities High Low 50 2 n 0.0 –0.2 15 C

Φ 25 C –0.4 –0.6 B. medius G. signata G. capito F. lapilum R. decemdigitatus F. varium K. stewarti R. whero F. malcolmi

Fig. 4.7 Structural and functional properties of haemoglobins of nine species of triplefin (Blenniidae) in relation to their habitat (preferential depths). The species living in the unstable environment (i.e. rock pools) in the upper layers have a larger number of preferential cathodic haemoglobin components (those migrating towards the negatively charged cathode having isoelectric points, pI, >6.25), with higher oxygen affinities and lower cooperativity (expressed by the Hill coefficient, n50) as well as a much lower pH sensitivity of oxygen binding (expressed by the Bohr coefficient

D log P f= 50 , DpH where P50 is the partial pressure of oxygen at half saturation). Those species living in the deeper water, with less fluctuations in temperature and oxygen availability, had fewer haemoglobins. These were preferentially anodic (pI < 6.25), with a larger potential for allosteric regulation. The genera examined are Bellapiscis (B.), Grahamina (G.), Fosterygion (F.), Ruanoho (R.) and Karalepis (K.). (Source: Brix et al. 1999.)

ern and western Norway, and 70% in the Kattegat and coastal groups as well as between populations (Frydenberg et al. 1965). A similar, though less of coastal cod. Karpov and Novikov (1980) reported clear, cline can be seen along the North American on the functional properties of the HbI11, HbI12 and east coast (Sick 1965b). According to these publica- HbI22 components. They suggested that the HbI22 tions, genetic differences exist between Arctic cod molecule has the highest oxygen affinity and is the Physiology of Living in Water 89 most efficient oxygen carrier at low temperatures, Coping with low oxygen availability while the HbI11 molecule has the highest affinity at about 20°C. Brix et al. (1998a, submitted) and Environmental hypoxia causes a fish to maximize Pörtner et al. (2000b) clearly demonstrate that oxygen transport by cellular adjustments in the

HbI22 is better fitted to cold temperatures than red blood cells combined with a reduced heart- HbI11 because it is able to transport more oxygen beat (bradycardia), an increase in cardiac stroke from the environment to the tissues. This could volume, increased peripheral resistance, and result in a higher growth rate, a suggestion sup- enhanced efficiency of gas exchange linked to ported by length and weight data. However, while increased lamellar recruitment in the gills not excluding this possibility, it could also be a re- (Satchell 1971; Booth 1978). Hypoxia causes the sult of higher oxygen demand in response to cold following: adaptation, compensating for some of the energy 1 reduction of ATP (GTP) production in the cells, loss caused by proton leakage in the mitochondria which in turn changes the Donnan equilibrium,

(see above). In both cases the HbI22 phenotype ap- the passive distribution of solutes over a semiper- pears better fitted for life in cold environments. In meable membrane, causing a decrease in the the temperature range 8–12°C, where maximal concentration of protons and thus an increase in mean growth rates for all phenotypes have been intracellular pH; reported, there are no differences with respect to 2 increase in ventilation, causing increased oxygen-binding properties. Oxygen binding of all plasma pH and thus intracellular pH; phenotypes is very sensitive to pH. Any tempera- 3 increase in Hb/HbO2 ratio (chronic hypoxia), ture change would thus greatly affect oxygen causing an increase in intracellular pH. affinities by indirectly changing pH (Brix et al. The reduction in ATP (GTP) and the increase in 1981). Brix et al. (submitted) further showed that intracellular pH increases oxygen affinity, which the heterozygote, HbI12, changed the concentra- thus safeguards oxygen uptake (Weber et al. 1976). tion of haemoglobin components during long- When oxygen availability in water becomes lim- term acclimation to either 4°C or 12°C, achieving ited fish have to breathe air. similar oxygen-binding properties as HbI11 in the warmer water and vice versa. This clearly suggest Air breathing Various groups of fish have solved that selection is very important for the evolution the problem of air breathing in different ways, of cod. These results are supported by the work of ranging from simple modification of the gills that Árnason et al. (1998), who found that the stock prevent them from collapsing in air to using the from both the Baltic and Barents Sea are more mouth, special parts of the gut, the swimbladder, related to the North Atlantic stock than with each or even the development of lungs. The role of the other, supporting the hypothesis of a transatlantic gills in oxygen uptake is reduced but they are still flow of genes (Árnason et al. 1992). More detail on the most important site for carbon dioxide elimi- the genetic structure of North Atlantic cod stocks nation. The air-breathing organs in bimodally is given by Ward (Chapter 9, this volume). breathing fish are well vascularized with an effi- Temperature not only influences the metabo- cient blood supply. In species using modified gills, lism of fish but also oxygen availability. At high mouth or opercular cavities for air breathing, the temperatures the oxygen concentration in water air-breathing organs are parallel with the gills and will be markedly reduced, which makes it very dif- the blood enters directly into the systemic circula- ficult for fish to meet the extra oxygen demand. We tion. For most other bimodal breathers, blood from refer to this condition as environmental hypoxia, the air-breathing organ passes the gills before and fish have to make significant circulatory and entering the systemic circulation. The lungfish, respiratory adjustments in order to extract suffi- Dipnoi, have separated the pulmonary and cient oxygen from the water. branchial circulation allowing them, unlike most other air-breathing fishes, to respire with their 90 Chapter 4 lungs and gills simultaneously (Johansen et al. 1990). The ability to triturate, contain, and retain 1968). The tetrapods evolved from this ancient plant material will therefore have a major impact group of fish (see Gill and Mooi, Chapter 2, this on the structure and function of the alimentary volume). tract, in particular the possible role of gastro- As water dries up and becomes more and more intestinal microbes in digestion. The material in muddy, the African lungfish (Protopterus) makes a this section is complementary to the chapter bottle-shaped burrow lined by mucus secreted by on growth by Jobling (Chapter 5, this the skin to form a cocoon and the fish becomes tor- volume) and the chapter on the behavioural pid. The nares becomes plugged with mud and the ecology of feeding by Mittelbach (Chapter 11, this fish breathes air through its mouth about once an volume). hour. The lungfish may remain aestivated for more than 6 months and may survive for years in this 4.6.2 Carnivorous fish condition. During aestivation, when ammonia ex- cretion over the gills is impossible, urea is formed As in other vertebrates the alimentary tract in fish in the liver and accumulates in the blood. Other is divided into the mouth and buccal cavity, the species, like the western Australian Lepidogala- pharynx, the oesophagus, the stomach, the intes- xias salamandroides and the New Zealand mud- tine with the pyloric caeca and related organs minnows (Neochanna), also aestivate during (liver, gallbladder and pancreas), and the rectum periods of drought. and anus. The alimentary tract is lined by an inner epithelium called the mucosa, underneath which is the submucosa, a layer of circular and longitudi- 4.6 DIGESTION AND nal muscles called the muscularis, and the serosa. ABSORPTION The thickness of the different layers varies in the 4.6.1 Structure and function of the different parts of the alimentary tract. alimentary tract Ingested food is sometimes broken down mechanically in the mouth and pharynx. In the Our knowledge of digestive physiology is based oesophagus, which is highly distensible, the food largely on studies conducted on carnivorous is lubricated by mucus secreted from the mucosa Northern Hemisphere species (Kapoor et al. 1975; before entering the stomach. The stomach in Fänge and Grove 1979; Helpher 1988; Lovell 1989), fishes can be a straight tube, U-shaped, or Y-shaped most of which are freshwater or diadromous taxa with a gastric caecum (Fänge and Grove 1979; such as salmonids (Christiansen and Klungsøyr Helpher 1988; Stevens 1988). Distension of the 1987). Thus, the overall morphology and enzyme stomach activates a cholinergic response, which complement of the alimentary tract appears to be triggers secretion from the gastric mucosa of hy- more or less ‘hard-wired’ to the ‘normal’ diet of the drochloric acid, which decreases pH, and protease fish (Jobling 1998). However, many fish in the trop- enzymes such as pepsin that have a pH optimum of ics and the Southern Hemisphere are herbivorous, 2–4. In the foregut, the caecal tissue or the pancre- a fact that has been very little considered in the atic tissue, which commonly envelops the caeca, physiological literature (Clements 1997). The low produces a secretion with a pH of 7–9 that contains nutritional value of diets high in fibre requires the trypsin, an enzyme that digests protein at alkaline ingestion and processing of a large volume of food pH. The pancreatic tissue is also the primary site (Stevens 1988). The extra time necessary for for the production of lipase and amylase, which microbial digestion of this refractory material re- digest fat and carbohydrate. Lipase activity has quires the retention of food within the gut for also been found in the caecal tissue and in the extended periods of time, and the ability of herbiv- upper part of the intestine (Chesley 1934). The orous fish to triturate ingested food influences the pyloric caeca may also secrete fluids that serve rate and efficiency of fermentation (Bjorndal et al. to buffer intestinal contents (Montgomery and Physiology of Living in Water 91

Pollak 1988). A muscular valve or sphincter, a fold other species by the ingestion of infected food or in the mucous membrane, controls the flow of food water. Some species seem to have evolved specific into the intestine (Stevens 1988). Assimilation of mechanisms for the retention and transfer of digested food also takes place in the intestine. endosymbionts. The parrotfishes, however, do not The behaviour and ecology of predatory fish is contain an obvious microbiota (Clements 1991), described in detail in Juanes et al. (Chapter 12, this probably due to the slurry of calcium carbonate in volume). the alimentary tract causing unfavourable condi- tions for the development of large populations of 4.6.3 Herbivorous fish microorganisms (Clements 1997). Most marine herbivorous fish belong to the order Perciformes (Horn 1989; Choat 1991). They have 4.7 BIOLUMINESCENCE typically a small gape for taking rapid small bites, 4.7.1 From oxygen defence to and a row of small closely spaced mandibular teeth deep-sea communication that serve to crop plant material or scrape it from the substratum. In some fishes these teeth have Bioluminescence, the emission of ecologically become fused into a parrot-like beak (Scaridae, functional light by living organisms, emerged Odacidae; Clements and Bellwood 1988). Some independently during evolution on several occa- herbivorous fish have developed pharyngal teeth sions. The substrates of the luminous reactions, on the modified fifth gill arch, which does not luciferins, are suggested to be the evolutionary carry gills (Helpher 1988; Horn 1989, 1992). Buccal core of most systems. The luciferases, the en- teeth, which prevent the escape of prey, are not zymes catalysing the photogenic oxidation of important in herbivorous fish. Some fish, like luciferin, thus serve to optimize the expression the surgeonfish (Acanthurus) and the girellids of the endogenous chemiluminescent properties (Ctenochaetus), have developed gizzard-like of luciferin. It has been suggested that the primary stomachs similar to those in birds that serve to function of coelenterazine, a luciferin with strong break down the cell walls of bacteria, blue-green antioxidative properties that occurs in many algae, diatoms and filamentous red and green marine bioluminescent groups, was originally the macroalgae, which are ingested along with some detoxification of deleterious oxygen derivatives sedimentary material (Payne 1978; Lobel 1981; such as the superoxide anion or peroxides (Rees Horn 1992). The stomach may be absent in many et al. 1998). When marine organisms began colo- herbivorous fishes like cyprinids, odacids and nizing deeper layers of the oceans, where exposure scarids. These fish typically have a pharyngeal to oxidative stress is considerably lowered because apparatus for processing food before it enters the of reduced light irradiance and lower oxygen intestine (Fänge and Grove 1979; Stevens 1988). levels, the strength of selection for antioxidative Most species containing high concentrations of defence mechanisms decreased and their light- short-chain fatty acids (SCFA) in the gut lack a emitting function evolved. The endogenous pro- mechanism for mechanical triturating ingested duction of reactive oxygen species would decrease material (Clements 1997). Specialized fermenta- in parallel with the decrease in metabolic activity tion chambers have been found in a few species, at increasing depths, and mechanisms for harness- but high concentrations of SCFA in the gut of fish ing the chemiluminescence of coelenterazine in lacking these chambers suggests that material is specialized organs could have developed, while the retained in the intestine long enough for fermenta- beneficial antioxidative properties were main- tion to take place. It appears that gastrointestinal tained in other tissues. microorganisms are transmitted in some species Some species use lights as lures, such as barbells by the ingestion of infected faecal material, as and fishing rods with luminous tips, whilst others occurs in herbivorous reptiles (Troyer 1984), and in have light organs in the mouth or even headlights 92 Chapter 4 to illuminate their prey. For the great majority filter transmitting at 485nm, light then leaves the of these fish, photophores are used for signal- photophore in a particular pattern that differs in ling other members of the same species, or for intensity at different angles (Bone et al. 1996). camouflage. Most bioluminescent fish are self- luminescent, although a substantial minority of bioluminescent teleosts produce light that is 4.8 CONCLUSIONS due to symbiotic luminous bacteria housed in elaborate light organs. In this chapter I have reviewed some of the most important physiological mechanisms that fish 4.7.2 Symbiotic luminescence use to cope with the four major physical and chemical constraints imposed on life in the The majority of symbiotically bioluminescent aquatic environment. fishes, in ten families in five orders, harbour The high density of water reduces the effects of common free-living species of marine luminous gravity and enables fish to remain suspended in the bacteria: Photobacterium phosphoreum, P. leio water column using a minimum amount of energy. gnathi and P. fischeri (= Vibrio fischeri). Others, as- Fish cope with buoyancy problems by applying sociated with the beryciform family Anomalopi- either dynamic or static lift or a combination of dae and nine families in the lophiiform suborder both. Lift forces are generated by the caudal and Ceratioidei, have apparently obligate symbionts pectoral fins acting as lifting foils at about the cen- that have recently been identified by small subunit tre of gravity of the fish and depend on a minimum (16S) rRNA analysis as new groups within the cruising speed to prevent the fish sinking. Static genus Vibrio (Haygood 1993). Fish with symbiotic lift is obtained by the use of lipids and/or gas. Since luminescence have normally not more than one or the volume of gas is inversely proportional to the two, or a maximum of four, light organs probably ambient pressure, which increases by 101kPa for due to problems of infection and maintenance. each 10m of depth, the swimbladder must have The same organ may serve several purposes, mechanisms for efficient filling and emptying at such as illuminating prey, schooling and sexual different depths. Gas can be released either via the communication. pneumatic duct into the gut whilst ascending/ diving (physostomes), or into the circulation via 4.7.3 Self-luminescence the oval chamber (physostomes/physoclists). The low compressibility of water influences Fish with self-luminescence often have many the swimming performance of fish, which have to photophores, which allows them to use different push water aside to move through it. For this pur- ones for different purposes. Within the upper pose they have developed different motor systems, 1000m of the ocean, down-welling light will reduced the drag forces and the cost of locomotion, silhouette the fish when looked at from below. and increased the efficiency of respiratory and Fish with photophores can make themselves cardiovascular adaptations during swimming. invisible when viewed obliquely or directly from The most important characteristic of water is below by emitting light that matches the angular that it is an almost universal solvent, containing distribution and wavelength as well as the inten- complex mixtures of salts, organic compounds and sity of the ambient background light. The pho- gases. Many of these are essential for life and are tophores, which are richly innervated by the taken up by fish using specialized organs in com- autonomic nervous system, are arranged in ven- plicated processes of ion regulation, osmoregula- trally directed tubes along the lower part of the tion, gas exchange and acid–base balance, or via fish. The light is produced in a dorsal chamber, digestion of food. The osmolality of fresh water is lined with black pigment apart from a series of <10mosmol, and sea water falls within the range small ventral windows. After passing a colour 800–1300mosmol. The freshwater teleosts have Physiology of Living in Water 93

-1 blood osmolalities of about 260–330mosmolkg . REFERENCES Thus these fish will be subjected to a passive influx of water and efflux of ions. In marine fishes, which Árnason, E., Palsson, S. and Arason, A. (1992) Gene have blood osmolalities in the range 370–480 flow and lack of population differentiation in Atlantic -1 mosmolkg , these fluxes will be reversed. At cod, Gadus morhua L., from Iceland, and comparison subzero temperatures, marine fish may produce of cod from Norway and Newfoundland. Journal of macromolecular ‘antifreeze’ compounds in their Fish Biology 40, 751–70. blood serum. The antifreeze consists of glyco- Árnason, E., Petersen, P.H. and Palsson, S. (1998) Mito- proteins or proteins, which coat and isolate ice chondrial cytochrome b DNA sequence variation of Gadus morhua crystals forming in the blood. Atlantic cod, , from the Baltic and the White Seas. Hereditas 129, 37–43. Water contains very little oxygen compared Bennion, G.R. (1968) The control of the function of the with air. Fish have therefore evolved a unidirec- heart in teleost fish. MS thesis, University of British tional flow of water over their gills to maximize Columbia, Vancouver. oxygen uptake. Cutaneous respiration may be Berg, T. and Steen, J.B. (1965) Physiological mechan- occasionally of some significance either when isms for aerial respiration in the eel. Comparative oxygen demand is very low, as in some Antarctic Biochemistry and Physiology 15, 469–84. Binotti, I., Giovenco, S., Giardina, B., Antonini, E., fishes, or when the diffusion distance is very short, Brunori, M. and Wyman, J. (1971) Studies on the func- as for fish larvae. The main tasks of the circulatory tional properties of fish haemoglobins. II. The oxygen system are to transport respiratory gases, nutri- equilibrium of the isolated haemoglobin components ents, metabolic waste products, endocrine factors from trout blood. Archives of Biochemistry and and heat. To ensure an adequate supply of oxygen Biophysics 142, 274–80. to all parts of the organism, oxygen is transported Bjorndal, K.A., Bolten, A.B. and Moore, J.E. (1990) Diges- by haemoglobin. Different haemoglobins have tive fermentation in herbivores: effect of food particle size. Physiological Zoology 63, 710–21. evolved and have acquired special features to meet Bone, Q. (1978) Muscles and locomotion. In: T.J. special needs. Both the number of these so-called Shuttleworth (ed.) Physiology of Elasmobranch isohaemoglobins and their functional heterogene- Fishes. Berlin: Springer Verlag, pp. 99–142. ity is related to the constancy of physicochemical Bone, Q., Marshall, N.B. and Blaxter, J.H.S. (eds) (1996) conditions in the environment. Biology of Fishes, 2nd edn. London: Chapman & Hall. Temperature is probably the most important Booth, J.H. (1978) The distribution of blood flow in the environmental factor influencing the life of fish. gills of fish: application of a new technique to rainbow trout. Journal of Experimental Biology 73, 119–29. Oxygen transport may be optimized to different Brett, J.R. (1975) The swimming energetics of salmon. temperatures by changing the concentrations of Scientific American 212, 80–5. isohaemoglobins, as seen for the Atlantic cod. At Brix, O., Lykkeboe, G. and Johansen, K. (1981) The sig- high temperatures, the oxygen concentration in nificance of the linkage between the Bohr and Haldane water will be markedly reduced, which makes effects in cephalopod bloods. Respiration Physiology it very difficult for fish to meet the extra oxygen 44, 177–86. Brix, O., Forås, E. and Strand, I. (1998a) Genetic variation demand. Fish thus have to make significant cir- and functional properties of Atlantic cod hemoglobins: culatory and respiratory adjustments in order to introducing a modified tonometric method for study- extract sufficient oxygen from the water. ing fragile hemoglobins. Comparative Biochemistry The transparency of water is very poor. Since and Physiology A 119, 575–83. most food production is within the photic zone, Brix, O., Clements, K.D. and Wells, R.M.G. (1998b) An most fish reside in this region, where prey capture ecophysiological interpretation of hemoglobin multi- depends on good visibility and well-developed plicity in three herbivorous marine teleost species from New Zealand. Comparative Biochemistry and eyes. In the dark zone of the oceans, however, fish Physiology A 121, 189–95. that communicate and capture prey by producing Brix, O., Clements, K.D. and Wells, R.M.G. (1999) their own light using photophores. Haemoglobin components and oxygen transport in relation to habitat distribution in triplefin fishes (F. 94 Chapter 4

Tripterygiidae). Journal of Comparative Physiology B Dobbs, G.H., Lin, Y. and DeVries, A.L. (1974) Aglomeru- 169, 323–34. larism in Antarctic fish. Science 185, 793–4. Brix, O., Thorkildsen, S. and Colosimo, A. (submitted) Duman, J.G. and DeVries, A.L. (1974a) Freezing resis- Temperature acclimation modulates the oxygen bind- tance in winter flounder Pseudopleuronectes ameri- ing properties of the Atlantic cod genotypes HbI*1/1, canus. Nature 247, 237–8. HbI*1/2 and HbI*2/2 by changing the concentrations Duman, J.G. and DeVries, A.L. (1974b) The effects of of their major haemoglobin components. temperature and photoperiod on antifreeze production Brunori, M., Coletta, M. and Giardina, B. (1985) Oxygen in cold water fishes. Journal of Experimental Zoology carrier proteins. In: P. Harrison (ed.) MetalloProteins, 190, 89–97. 2. London: MacMillan, pp. 263–331. Eaton, W.A. (1982) The relationship between coding Chadwick, T.D. and Wright, P.A. (1999) Nitrogen excre- sequences and functions in haemoglobin. Nature 284, tion and expression of urea cycle enzymes in the 183–5. Atlantic cod (Gadus morhua L.): a comparison of Fago, A., Carratore, V., DiPrisco, G., Fuerlein, R.J., early life stages with adults. Journal of Experimental Sottrup Jensen, L. and Weber, R.E. (1995) The Biology 202, 2653–62. cathodic hemoglobin of Anguilla anguilla: amino acid Chesley, L.C. (1934) The concentrations of proteases, sequence and oxygen equilibria of a reverse Bohr effect amylase and lipase in certain marine fishes. Biological hemoglobin with high affinity and high phosphate Bulletin 6, 133–44. sensitivity. Journal of Biological Chemistry 270, Choat, J.H. (1991) The biology of herbivorous fishes 18897–902. on coral reefs. In: P.F. Sale (ed.) The Ecology of Fishes Fago, A., Wells, R.M.G. and Weber, R.E. (1997a) on Coral Reefs. San Diego: Academic Press, pp. Temperature-dependent enthalpy of oxygenation 120–55. in Antarctic fish hemoglobins. Comparative Christiansen, D.C. and Klungsøyr, L. (1987) Metabolic Biochemistry and Physiology B 118, 319–26. utilization of nutrients and the effects of insulin in Fago, A., Bendixen, E., Malte, H. and Weber, R.E. (1997b) fish. Comparative Biochemistry and Physiology B 88, The anodic hemoglobin of Anguilla anguilla. Journal 701–11. of Biological Chemistry 272, 15628–35. Clements, K.D. (1991) Gut microorganisms of surgeon- Fänge, R. and Grove, D. (1979) Digestion. In: W.S. Hoar, fishes (family Acanthuridae). PhD thesis, James D.J. Randall and J.R. Brett (eds) Fish Physiology, Vol. Cook University of North Queensland, Townsville, VIII. New York: Academic Press, pp. 161–260. Australia. Freadman, M.A. (1979) Role partitioning of swim- Clements, K.D. (1997) Fermentation and gastrointesti- ming musculature of striped bass, Morone saxatilis nal microorganisms in fishes. In: R.I. Mackie and Walbaum and bluefish, Pomotomus saltatrix L. B.A. White (eds) Gastrointestinal Microbiology. Vol. 1. Journal of Fish Biology 15, 417–23. Gastrointestinal Ecosystems and Fermentations. Frydenberg, O., Møller, D., Nævdal, G. and Sick, K. New York: Chapman & Hall, pp. 156–98. (1965) Hemoglobin polymorphism in Norwegian cod Clements, K.D. and Bellwood, D.R. (1988) A comparison populations. Hereditas 53, 257–71. of the feeding mechanisms of two herbivorous labroid Fyhn, H.J. (1993) Multiple functions of free amino acids fishes, the temperate Odax pullus and the tropical during embryogenesis in marine fishes. In: B.T. Walter Scarus rubroviolaceus. Australian Journal of Marine and H.J. Fyhn (eds) Physiological and Biochemical and Freshwater Research 39, 87–107. Aspects of Fish Development. Bergen: University of Cocca, E., Ratnayake-Lecamwasam, M., Parker, S.K. Bergen, pp. 299–308. et al. (1997) Do the hemoglobinless icefishes have glo- Fyhn, H.J., Finn, R.N., Reith, M. and Norberg, B. (1999) bin genes? Comparative Biochemistry and Physiology Yolk protein hydrolysis and oocyte free amino acids as A 118, 1027–30. key features in the adaptive evolution of teleost fishes DeVries, A.L. and Wohlschlag, D.E. (1969) Freezing resis- to seawater. Sarsia 84, 451–6. tance in some Antarctic fishes. Science 163, 1073–5. Fyhn, U.E.H., Fyhn, H.J., Davis, B.J., Powers, D.A., Fink, DiPrisco, G. and Tamburrini, M. (1992) The hemoglobins W.L. and Garlick, R.L. (1979) Hemoglobin heterogene- of marine and freshwater fish: the search for correla- ity in Amazonian fishes. Comparative Biochemistry tions with physiological adaptation. Comparative and Physiology A 62, 39–66. Biochemistry and Physiology B 102, 661–71. Guppy, M. and Hochachka, P.W. (1978) Skipjack tuna DiPrisco, G., D’Avino, R., Carmadella, L., Caruso, C., white muscle: a blueprint for the integration of aerobic Romano, M. and Rutigliano, B. (1990) Structure and and anaerobic carbohydrate metabolism. In: G.D. function of hemoglobin in Antarctic fishes and evolu- Sharp and A.E. Dizon (eds) The Physiological Ecology tionary implications. Polar Biology 10, 269–74. of Tunas. New York: Academic Press, pp. 175–81. Physiology of Living in Water 95

Haygood, M.G. (1993) Light organ symbioses in fishes. development of the vertebrate kidney in relation to Critical Reviews in Microbiology 19, 191–216. habitat. Biological Bulletin Woods Hole 59, 135–53. Heisler, N. (1984) Acid–base regulation in fishes. In: W.S. Marshall, N.B. (ed.) (1996) The Life of Fishes. New York: Hoar and D.J. Randall (eds) Fish Physiology, X, Part A. Universe Books. New York: Academic Press, pp. 315–401. McInerney, J.E. (1974) Renal sodium reabsorption in the Helpher, B. (ed.) (1988) Nutrition of Pond Fishes. hagfish, Eptatretus stouti. Comparative Biochemistry Cambridge: Cambridge University Press. and Physiology A 49, 273–80. Horn, M.H. (1972) The amount of space available for Montgomery, W.L. and Pollak, P.E. (1988) Gut anatomy marine and freshwater fishes. National Oceanic and and pH in a Red Sea surgeonfish, Acanthurus nigrofus- Atmospheric Administration, Fishery Bulletin 70, cus. Marine Ecology Progress Series 44, 7–13. 1295–7. Murad, A. and Houston, A. (1991) Haemoglobin iso- Horn, M.H. (1989) Biology of marine herbivorous fishes. morph abundances in splenectomized rainbow trout, Oceanography and Marine Biology Annual Review Oncorhynchus mykiss (Walbaum). Journal of Fish 27, 167–272. Biology 38, 641–51. Horn, M.H. (1992) Herbivorous fishes: feeding and Nakano, T. and Tomlinson, N. (1967) Catecholamine and digestive mechanisms. In: D.M. John, S.J. Hawkins carbohydrate concentrations in rainbow trout (Salmo and J.H. Price (eds) Plant–Animal Interactions in gairdneri) in relation to physical disturbance. Journal the Marine Benthos. Oxford: Clarendon Press, pp. of the Fisheries Research Board of Canada 24, 339–62. 1701–15. Houston, A.H. and Gingras-Bedard, J.H. (1994) Variable Nelson, J.S. (ed.) (1994) Fishes of the World, 3rd edn. New versus constant temperature acclimation regimes: York: John Wiley & Sons. effects on hemoglobin isomorph profile in goldfish, Payne, A.I. (1978) Gut pH and digestive strategies in estu- Carassius auratus. Fish Physiology and Biochemistry arine grey mullet (Mugilidae) and tilapia (Cichlidae). 13, 445–50. Journal of Fish Biology 13, 627–9. Jobling, M. (ed.) (1998) Environmental Biology of Fishes. Pörtner, H.O., Hardewig, I., Sartoris, F.J. and van Dijk, London: Chapman & Hall. P.L.M. (1998) Energetic aspects of cold adaptation: Johansen, K., Lenfant, C., Schmidt-Nielsen, K. and critical temperatures in metabolic, ionic and acid–base Petersen, J. (1968) Z. vergl. Physiologic 61, 137–63. regulation? In: H.O. Pörtner and R. Playle (eds) Cold Kapoor, B.G., Smit, H. and Verighina, I.A. (1975) The Ocean Physiology. Cambridge: Cambridge University alimentary canal and digestion in teleosts. Advances Press, pp. 88–120. in Marine Biology 13, 109–230. Pörtner, H.O., van Dijk, P.L.M., Hardewig, I. and Karpov, A.K. and Novikov, G.G. (1980) The hemoglobin Sommer, A. (2000a) Levels of metabolic cold adapta- aloforms in cod (Gadus morhua L), their functional tion: tradeoffs in eurythermal and stenothermal characteristics and distribution in the populations. ectotherms. In: W. Davison and C. Howard Williams Journal of Ichthyology 20, 45–50. (eds) Antarctic Ecosystems: Models for Wider Kirsch, R., Meens, R. and Meister, M.F. (1981) Osmoregu- Ecological Understanding. Christchurch New lation chez les teleostéens marins: rôle des branchies et Zealand: Caxton Press, pp. 109–22. du tube digestif. Bulletin de la Société Zoologique de Pörtner, H.O., Berdal, B., Blust, R. et al. (2000b) Growth France 106, 31–6. performance, fecundity and recruitment in marine Lahlou, B. (ed.) (1980) Epithelial Transport in the Lower fish: understanding climate effects based on physiol- Vertebrates. Cambridge: Cambridge University Press. ogical and ecological studies in Atlantic cod (Gadus Laurent, P. (1984) Gill internal morphology. In: W.S. Hoar morhua) and common eelpout (Zoarces viviparus). and D.J. Randall (eds) Fish Physiology, X, Part A. New In: Proceedings of the 4th European Conference on York: Academic Press, pp. 73–183. Marine Science and Technology, Vol. I. Marine Lobel, P.S. (1981) Trophic biology of herbivorous Processes, Ecosystems and Interactions, pp. 190–6. reef fishes: alimentary pH and digestive capabilities. Potts, W.T.W. (1976) Ion transport and osmoregulation in Journal of Fish Biology 19, 365–97. marine fish. In: P.S. Davies (ed.) Perspectives in Experi- Lovell, T. (ed.) (1989) Nutrition and Feeding of Fish. New mental Biology. Oxford: Pergamon Press, pp. 65–75. York: Van Nostrand Reinhold. Powers, D.A. (1974) Structure, function, and molecular Mangor-Jensen, A. and Adoff, G.R. (1987) Drinking ecology of fish hemoglobins. Annals of the New York activity of the newly hatched larvae of the cod Academy of Sciences 241, 472–89. Gadus morhua L. Fish Physiology and Biochemistry 3, Powers, D.A. (1980) Molecular ecology of teleost fish 99–103. hemoglobins: strategies for adapting to changing Marshall, E.K. and Smith, H.W. (1930) The glomerular environments. American Zoologist 20, 139–62. 96 Chapter 4

Randall, D.J. (1970) The circulatory system. In: W.S. Hoar Thorsen, A. and Fyhn, H.J. (1996) Final oocyte matura- and D.J. Randall (eds) Fish Physiology, 4. New York: tion in vivo and in vitro in marine fishes with pelagic Academic Press, pp. 133–72. eggs: yolk protein hydrolysis and free amino acid Rees, J.F., DeWergifosse, B., Noiset, O., Dubuisson, M., content. Journal of Fish Biology 48, 1195–209. Janssens, B. and Thompson, E.M. (1998) The origins Troyer, T.J. (1984) Microbes, herbivory and the evolution of marine bioluminescence: turning oxygen defence of social behaviour. Journal of Theoretical Biology 106, mechanisms into deep-sea communication tools. 157–69. Journal of Experimental Biology 201, 1211–21. Val, A.L., de Almeida-Val, V.M.F. and Affonso, E.G. (1990) Rønnestad, I., Robertson, R. and Fyhn, H.J. (1996) Adaptive features of Amazon fishes: hemoglobins, Chemical composition of tropical marine fish eggs hematology, intraerythrocytic phosphates and whole related to phylogeny: protein and free amino acids. blood Bohr effect of Pterygoplichthys multiradiatus In: D. MacKinlay and M. Eldridge (eds) The Fish Egg: (Siluriformes). Comparative Biochemistry and Its Biology and Culture. San Francisco, CA: The Physiology B 97, 435–40. American Fisheries Society, San Francisco State Verheyen, E., Blust, R. and Decleir, W. (1986) Hemoglo- University, pp. 81–4. bin heterogeneity and the oxygen affinity of the Samuelsen, E.N., Imsland, A.K. and Brix, O. (1999) hemolysate of some Victoria cichlids. Comparative Oxygen binding properties of three different hemo- Biochemistry and Physiology A 84, 315–18. globin genotypes in turbot (Scophthalmus maximus Wardle, C.S. and Reid, A. (1977) The application of Rafinesque): effect of temperature and pH. Fish large amplitude elongated body theory to measure Physiology and Biochemistry 20, 135–41. swimming power in fish. In: J.H. Steel (ed.) Fisheries Satchell, G.H. (1965) Blood flow through the caudal vein Mathematics. London: Academic Press, pp. 171–91. of elasmobranch fish. Australian Journal of Science Webb, P.W. (1975) Acceleration performance of rainbow 27, 241–2. trout (Salmo gairdneri) and green sunfish (Lepomis Satchell, G.H. (ed.) (1971) Circulation in Fishes. London: cyanellus). Journal of Experimental Biology 63, Cambridge University Press. 451–65. Schmidt-Nielsen, K. (ed.) (1997) Animal Physiology. Weber, R.E., Lykkeboe, G. and Johansen, K. (1976) Adaptation and Environment, 5th edn. Cambridge: Physiological properties of eel hemoglobin: hypoxic Cambridge University Press. acclimation, phosphate effects and multiplicity. Sick, K. (1965a) Hemoglobin polymorphism of cod in the Journal of Experimental Biology 64, 75–88. Baltic and the Danish belt sea. Hereditas 54, 19–48. Wells, R.G.M. and Jokumsen, A. (1982) Oxygen binding Sick, K. (1965b) Hemoglobin polymorphism of cod in the properties of hemoglobins from Antarctic fishes. Com- North Sea and the North Atlantic Ocean. Hereditas parative Biochemistry and Physiology B 71, 459–73. 54, 49–68. Wood, C.M. and Shelton, G. (1975) Physical and adrener- Steen, J.B. (ed.) (1971) Comparative Physiology of gic factors affecting systemic vascular resistance in the Respiratory Mechanisms. New York: Academic Press. rainbow trout: a comparison with branchial vascular Stevens, C.E. (ed.) (1988) Comparative Physiology of the resistance. Journal of Experimental Biology 63, Vertebrate Digestive System. Cambridge: Cambridge 505–23. University Press. Wood, S.C. and Lenfant, C. (1979) Oxygen transport and Terjesen, B.F., Mangor-Jensen, A. and Fyhn, H.J. oxygen delivery. In: S.C. Wood and C. Lenfant (eds) (1998) Ammonia dynamics in relation to hatching in Evolution of Respiratory Processes. A Comparative Atlantic halibut (Hippoglossus hippoglossus L.). Fish Approach. New York: Marcel Dekker, pp. 193–223. Physiology and Biochemistry 18, 189–201. Wyman, J. (1964) Linked functions and reciprocal effects Terjesen, B.F., Rønnestad, I., Norberg, B. and Anderson, in hemoglobin: a second look. Adv. Protein. Chem. P.M. (2000) Detection and basic properties of car- Anson. 19, 223–86. bamoyl phosphate synthetase III during teleost Zielinski, S. and Pörtner, H.O. (1996) Energy metabolism ontogeny: a case study in the Atlantic halibut and ATP free-energy change of the intertidal worm, (Hippoglossus hippoglossus L.). Comparative Sipunculus nudus, below a critical temperature. Biochemistry and Physiology B 126, 521–35. Journal of Comparative Physiology B 166, 492–500. 5 Environmental Factors and Rates of Development and Growth

MALCOLM JOBLING

5.1 INTRODUCTION a ‘critical weight’ and a ‘critical level of nutritional status, or fatness’, and now incorporates notions Information about fish age, development and about the way in which specific metabolic signals growth is a cornerstone in fishery research related to size and nutrition may be assessed and and management. Development describes the se- integrated within the brain and central nervous quencing of life-history stages and growth is a system (Robinson 1996; Kiess et al. 1999; Magni measure of change in size of either the whole body et al. 2000; Mystkowski and Schwartz 2000). Thus or some body part; growth rate is a measure of size there are close interrelationships between the change as a function of time. Growth depends availability of food, increase in size and the accu- upon the quantity and quality of food ingested, mulation of energy reserves, the timing of sexual with inadequate nutrition both retarding growth maturation and reproduction. These, in turn, re- and delaying developmental transitions, such as late to production, which is determined by the rate the timing of onset of sexual maturation. When of reproduction, the rate of growth of individuals food is limited the onset of maturation may be within the population, and the rate of mortality. delayed for months or years, until good feeding These functional rates determine population dy- conditions arise. In other words, the timing of namics over time, as well as structural elements sexual maturation appears to be more closely of the population, such as biomass, density and associated with size than age, leading to the con- size–frequency distribution, at any point in time. cept that maturation is initiated once a ‘critical As such, information about age and growth is ex- size’ has been attained. Further, during adult- tremely important in almost every aspect of fish- hood, fecundity or gamete production may be eries. The purpose of this chapter is to examine related to body size, and alterations in nutrition ways in which age and growth data can be collected can lead to either depression or enhancement of and analysed, and to provide a brief overview of the reproductive activity (see Hutchings, Chapter 7, environmental factors that may have an impact this volume). upon rates of development, growth and reproduc- Perhaps the most intriguing issue involved in tion of fish. understanding development, growth and sexual maturation is determining how the brain ‘knows’ when the body has reached the appropriate size for 5.2 TERMINOLOGY OF initiation of the physiological cascade that culmi- LIFE-HISTORY STAGES nates in the production of the first gametes. Over recent years, the concept of maturation at some The problem of terminology and sequencing of the ‘critical size’ has been extended to encompass both life-cycle intervals of fish is a long-standing one, 98 Chapter 5 and particularly so for the earliest part of the life history. A good terminology should be simple and linked to form and function. Difficulties arise in producing a terminology that meets the needs of workers representing different disciplines (e.g. de- velopmental biology, taxonomy, physiology, fish- eries biology and management) and encompasses the myriad patterns of development seen in fish. For example, some species of fish hatch in a well- developed state, especially where there is vivi- parity, ovoviviparity or a protracted incubation period within the egg; other species have short egg incubation times and the fish hatch at a much earlier state of development (Blaxter 1988; Balon 1990; Jobling 1995; Hutchings, Chapter 7, this volume). Fig. 5.1 Recently hatched, free embryos Balon (1975, 1984, 1990) proposed a saltatory (eleutheroembryos) of two fish species that differ in model of development in which the entire life his- developmental pattern. (a) The Atlantic cod, Gadus morhua, has small (1.1–1.9mm diameter) pelagic eggs tory, from the fertilization of the egg until death, with a relatively short incubation time (9–25 days) and can be divided into five periods: embryonic, larval, the fish are small (3.5–4.5mm) at hatch. (b) Atlantic juvenile, adult and senescent. The model is termed salmon, Salmo salar, produce large (5–6mm diameter) ‘saltatory’ because the periods are separated by demersal eggs which hatch to give large (15–25mm) major thresholds involving a rapid transition to a offspring after protracted incubation (50–160 days). new ‘stabilized’ state. Each period may be divided (Source: from Jobling 1995.) into phases, as a convenient means of identifying different levels of morphological or physiological development. Steps, which are the shortest inter- free)(Fig. 5.1) and this continues until most of the vals of ontogeny, are separated by less dramatic yolk has been utilized and the fish begins to take thresholds, and are found within a phase. Meta- exogenous food. morphosis is a major threshold separating the lar- The larval period starts when the transition to val period from the juvenile or adult period. Early exogenous feeding has taken place. This period in the life of the fish developmental events are re- lasts until ossification of the axial skeleton. Some flected in rapid morphological and physiological embryonic organs persist, and there may be devel- changes, but in each successive period the develop- opment of organs which are later lost or replaced mental rate decreases until senescence and death. by others performing the same function. Examples The embryonic period, starting with fertiliza- are surface blood vessels on fins and filamentous tion of the egg, is characterized by endogenous nu- appendages used in gas exchange. The larval period trition. This may be from the yolk or, in the case may be divided into two phases: the protopterygio- of viviparous species, nutrition via special absorp- larval and the pterygiolarval phases. The first tive organs (Balon 1990). The embryonic period spans the interval between the first exogenous includes a cleavage phase, covering the time feeding and the start of fin differentiation. The between fertilization and the commencement of pterygiolarval phase terminates when the median organogenesis, and an embryonic phase within fin-fold is no longer visible. The larval period may which there is intense organogenesis within the be long, as in some species of eel (e.g. Anguilla egg membrane. The embryonic phase continues spp.), whereas in salmonids it is difficult to distin- until hatching, when the embryo enters the guish a distinct larval period. eleutheroembryonic phase (Greek eleutheros, The juvenile period begins once the fins are Rates of Development and Growth 99 fully differentiated, and the temporary larval or- deemed to be sufficiently significant to merit use gans have regressed or been replaced. The transi- as a criterion for demarcation of the transition tion from larva to young fish sometimes involves between developmental periods. Many workers, extensive changes from an unfish-like appearance however, consider the larval period to commence to one resembling the adult, i.e. there is a meta- at the point at which the fish hatch, and will refer morphosis. The juvenile period lasts until the fish to the newly hatched fish as a yolk-sac larva. mature and the first gametes are produced. Juve- Blaxter (1988) recommends that the terms ‘em- niles are characterized by their rapid growth, and bryo’ be used to cover the period from fertilization there can be a distinct juvenile body colour or pig- to the point of hatching, ‘larva’ to cover develop- mentation pattern. ment from hatching to metamorphosis, and ‘juve- The adult period starts with gonad maturation, nile’ for the period from metamorphosis to first culminating in the production of the first gametes. spawning. As a supplement to this, Kamler (1992) The adult period usually incorporates spawning provided a series of developmental divisions based runs or migrations (see Metcalfe et al., Chapter 8, upon the general sequence of growth patterns dis- this volume), specialized reproductive behaviour played during the early life of fish. An embryo (see Forsgren et al., Chapter 10, this volume), grows slowly from fertilization to hatching, and and changes in external morphology and colour. growth accelerates following the transition to a Spawning may be repeated for a number of years or yolk-sac larva. Larvae grow to reach the maximum seasons, as in iteroparous species, or may be per- size that their yolk reserves will support and then formed only once, as is the case with the semel- growth rate decreases, usually coincident with the parous species, such as some eels and salmonids start of exogenous feeding. Between the start of ex- (see Hutchings, Chapter 7, this volume). During ogenous feeding and final yolk resorption there is the adult period, available resources may be often an energy deficit, involving negative growth; directed more towards the development of the go- after final yolk absorption non-feeding larvae con- nads than towards somatic growth. Consequently, tinue to show weight loss and die from starvation, rates of growth are generally lower during this whereas feeding larvae show rapid growth. period than during the juvenile period. The senescent period covers a period of extre- mely slow growth during which reproductive activity may also be reduced. The period of senes- 5.3 DEVELOPMENT AND cence can last for several years, during which there GROWTH DURING EARLY is a gradual decline in the numbers of fertile ga- LIFE HISTORY metes produced (e.g. sturgeon, Acipenser sturio). Alternatively, the senescent period may last for no Rates of development and the timing of transi- more than a few days or weeks, during which time tional events from one life-history period to the the body undergoes rapid degenerative changes next are, to a certain extent, genetically deter- (e.g. Pacific salmon, Oncorhynchus spp.). mined and can be considered as species-specific Balon’s (1975, 1984) saltatory model and termi- characteristics. For example, large eggs, such as nology of ontogeny are holistic, and there are those of salmonids, take longer to hatch than small compelling arguments for their adoption (Balon eggs, and, in general, eggs from fish living at 1990, 1999). However, they have not been univer- high latitudes take longer to hatch than those from sally accepted due to a failure to highlight some the tropics (Fig. 5.2)(Kamler 1992; Jobling 1995; transitional events that many workers consider Rombough 1997). Environmental factors may, important. For example, according to Balon’s ter- however, have a major modifying influence on minology, the larval period is defined as starting rates of development. Ambient temperature, for when the fish undergoes the transition from en- example, is known to be one of the most potent dogenous to exogenous feeding. Hatching is not factors influencing both rates of development and 100 Chapter 5

100 A modification has been proposed in which tem- Marine perature may be measured from a point other than Freshwater 50 0°C. When T0 is defined as the new base tempera- ture the expression becomes: 30 20 ()TTD- 0 = constant. (5.2) 10 In this expression T0 is the so-called biological (c) zero, i.e. the incubation temperature at which, in 5 theory, the time to reach the given stage of devel-

3 opment would be infinite. T0 may be close to 0°C Incubation period (days) 2 for cold-water and temperate species, but will be higher in warm-water species. Thus, the introduc- (b) 1 tion of T0 would appear to make the hyperbolic (a) relationship applicable to the description of 0.5 0 5 10 15 20 25 30 35 developmental rates of a larger number of fish species, especially those which require warm Temperature ( C) water for egg incubation and embryo development Fig. 5.2 Influence of incubation temperature on the (e.g. Weltzien et al. 1999). Nevertheless, when the time to hatch for eggs of a range of marine and freshwater degree-day formula is applied to experimental data fish species. Note that, in addition to temperature, egg involving a wide range of temperatures it may not size also has a major influence on incubation period. The be found to be satisfactory. lines indicate estimated incubation times for eggs of (a) Numerous other mathematical expressions 0.5mm, (b) 2mm and (c) 5mm diameter, respectively. have been used to describe the curvilinear relation- (Source: from Rombough 1997.) ship between incubation time and temperature (Cossins and Bowler 1987; Kamler 1992), with the overall survival of the eggs and larvae (Blaxter both power law (i.e. D = aTb) and exponential (i.e. 1988, 1992; Kamler 1992; Rombough 1997). Rates D = ae-bT) equations having been widely used to de- of development are generally found to increase scribe the relationship (Kamler 1992; Rombough with increasing temperatures. However, incuba- 1997). However, both the power law and exponen- tion of eggs at high temperature may result in tial functions have the weakness that they do not increased incidence of malformations and take into account the fact that there are tempera- abnormalities in the embryo, or the death of the ture extremes outside of which egg incubation is eggs. Temperature can also influence size at hatch- impossible, and mortality will be complete. When ing, efficiency of yolk utilization, time to meta- data have been fitted to the power law function, morphosis, behaviour, rates of feeding and the constant b, which describes the inverse rela- metabolic demands. tionship between incubation time and tempera- The simplest relationship used to describe the ture, has usually been found to be within the range influence of incubation temperature on rates of de- 1–1.5. With the exponential function the value of b velopment is a hyperbolic function, which encap- will usually be 0.1–0.15. Since Q10 and the con- sulates the concept of ‘degree-days’. This states stant b in the exponential function relate to each b that the product of the incubation temperature (T other as in Q10 = 10 , these constants translate to in °C) and the time (D in days) required to reach any Q10 values of 3–4, meaning that the developmental particular stage of development, e.g. from fertiliza- rate roughly triples or quadruples with each 10°C tion to hatch, is constant: increase in temperature. Cold-water species ap- pear to be particularly responsive. Developmental T ¥ D = constant. (5.1) rate is also highly dependent upon egg size, with Rates of Development and Growth 101 the time to hatch at any given temperature being appear to be associated with year-to-year differ- almost an order of magnitude longer for large eggs ences in temperature in the period during spawn- than small (Fig. 5.2). Rombough (1997) summa- ing and early development (e.g. Brander 1979; rized data for various freshwater and marine Løken et al. 1994). Further, within year-classes a species (n = 171) in a regression equation that ac- protracted spawning season coupled to changes in counted for 84.4% of interspecific variability in water temperature over time may be associated the time for eggs to hatch: with meristic differences between those individu- als which hatch and develop at different times logD = 1.20 - 0.0494T + 0.203d (5.3) (Lindsey 1988). Thus, environmental influences on meristic counts may be superimposed upon ge- where D is incubation time (days), T is tempera- netically controlled responses (Lindsey 1988). For ture (°C) and d is egg diameter (mm). This equation example, Løken and Pedersen (1996) reported an gave an overall Q10 of 3.1. Rates of posthatch larval inverse relationship between incubation tempera- development appear to be less temperature sensi- ture and vertebral number in cod, Gadus morhua, tive than rates of prehatch embryonic develop- although vertebral numbers differed between the ment, Q10 values of about 2 seeming to apply to offspring of coastal cod and northeastern Atlantic rates of development of larvae following hatch cod when they were reared at the same tempera- (Rombough 1997). One should caution, however, ture. Meristic characters can be greatly modified that although it is possible to use Q10 to relate by the environment experienced during early de- temperature to developmental and physiological velopment. Later, but still quite early in ontogeny, processes, Q10 values do not remain constant over these characters become fixed and remain un- wide ranges of temperature. Thus, although Q10 changed thereafter. values can be useful for predictive purposes over Environmental factors, particularly tempera- narrow temperature ranges, extrapolations to tem- ture, probably influence meristic characters by perature extremes should not be attempted. differentially affecting the processes of growth In addition to influencing rates of development through body elongation and differentiation ex- of eggs, embryos and larvae, temperature may af- pressed as segment formation; temperature may fect the interactions between growth and differen- also have profound influences upon other develop- tiation. The best-known effect is that on meristic mental events such as muscle differentiation and characters. The number of serial structures such as the relative timing of organogenesis (Blaxter 1988, vertebrae, scales and gill rakers is labile and sus- 1992; Johnston 1993; Brooks and Johnston 1994; ceptible to environmental influence (Blaxter 1988; Nathanailides et al. 1995; Johnston et al. 1996, Lindsey 1988). There is also some evidence of a 1997). Environmental factors may also influence prefertilization influence on meristic characters, sex determination in several animal groups, in- i.e. the temperature experienced by the broodstock cluding fish (Baroiller et al. 1999; Baroiller and during gametogenesis may influence the meristic D’Cotta 2001). This phenomenon, known as characters of the offspring. environmental sex determination (ESD) (Janzen The number of vertebrae tends to be higher in and Paukstis 1991; Crews 1996; Shine 1999; see fish from polar and temperate waters than in their also Forsgren et al., Chapter 10, this volume), relatives from tropical waters. This phenomenon, means that the environment experienced by fish termed Jordan’s rule, is usually described in rela- during early development can influence pheno- tion to latitude, but may be attributable to the fact typic sex (Conover et al. 1992; Baroiller et al. 1995, that fish from different latitudes are usually sub- 1996, 1999; Craig et al. 1996; Patino et al. 1996; jected to different thermal environments during Römer and Beisenherz 1996; Strüssmann et al. early development (Lindsey 1988). In addition, 1996, 1997; Nomura et al. 1998). It is the effects of meristic counts of wild fish hatched in the same temperature on sex determination that have been place often vary between years, and the differences most studied in fish, although both salinity and 102 Chapter 5 xenobiotics are known to influence phenotypic ing. Numerical expressions of growth may be sex in some fish species (Baroiller et al. 1999; based on absolute changes in length or weight Baroiller and D’Cotta 2001). (absolute growth), or changes in length or weight Sex ratios that deviate significantly from the relative to the size of the fish (relative growth). theoretical 1:1 amongst groups of fish reared at dif- Length almost always increases with time, ferent temperatures could arise from differential whereas weight can either increase or decrease mortality of the sexes. Thus, deviations of the sex over a given time interval depending upon the ratio from the expected 1:1 amongst fish reared influences of the various factors that affect the under different temperature conditions cannot be deposition and mobilization of body materials. taken as providing conclusive evidence of ESD Measurements of growth in relation to time pro- (Strüssmann et al. 1997; Nomura et al. 1998). Con- vide an expression of growth rate. Growth in clusive evidence of ESD has, however, been ob- length can usually be modelled using an asymptot- tained by progeny testing of fish that have been ic curve which tapers off with increasing age (Fig. subjected to different temperature treatments, and 5.3). Growth in weight is usually sigmoidal, i.e. the by demonstration that temperature can induce a weight increment increases gradually up to an in- phenotypic sex change within gynogenetic mono- flection point from where it gradually decreases sex populations (Baroiller et al. 1995, 1996, 1999; again. Thus, growth rates are constantly changing, Nomura et al. 1998; Kitano et al. 1999; D’Cotta and the absolute growth increments will be differ- et al. 2001a,b). ent for different sizes of fish. Whether or not temperature has any influence on the phenotypic sex of individuals within a given species will depend upon the strength of genetic 120 sex determination, and when in development the temperature treatment is applied (Baroiller et al. 100 1995, 1996, 1999; Römer and Beisenherz 1996; Strüssmann et al. 1996, 1997). This is to be ex- 80 pected because determination of phenotypic sex in fish is also sensitive to sex steroid hormone treat- 60 ments only at particular stages of development (re- viewed by Purdom 1993 and Patino 1997). It is Length (cm) 40 unlikely that the periods of sensitivity to tempera- ture and sex steroids will differ, because the physi- ological effects of temperature treatment appear to 20 be mediated via actions on genes coding for P450 0 steroidogenic enzymes, such as aromatase, and sex 0 5 10 15 20 25 steroid hormone receptors (Crews 1996; Baroiller Age (years) et al. 1999; Kitano et al. 1999; Baroiller and D’Cotta 2001; D’Cotta et al. 2001a,b). Fig. 5.3 Length at age plot for female spiny dogfish,

Squalus acanthias. The parameters L• and k of the von Bertalanffy growth function are about 104cm and 0.106, 5.4 GROWTH MODELS respectively. Spiny dogfish are ovoviviparous, with AND EQUATIONS prolonged intrauterine incubation of large eggs. Extrapolation of the regression line gives an intersection on the y-axis above the origin, providing an indication of Growth equations are used to describe changes in intrauterine growth and an estimate of the length at the length or weight of a fish with respect to time, ‘birth’ (c.25cm), whereas the negative intersection on although the constants derived from such empiri- the x-axis gives an estimate of gestation time (c.2 years). cal equations may have no exact biological mean- (Source: data from Holden and Meadows 1962.) Rates of Development and Growth 103

A number of mathematical functions have been where Lt is length at time t, and t0 is the theoretical used to describe growth curves, including the ‘age’ of the fish at zero size.

Gompertz, the logistic, and a range of straight-line The two constants L• and k can be estimated and exponential approximations (Beverton and from measurements of fish length at known fish Holt 1957; Ricker 1979; Weatherley and Gill 1987; ages (Gulland 1969; Bagenal and Tesch 1978; Prein Prein et al. 1993; Elliott 1994). Life-history pat- et al. 1993). Before personal computers became terns of fish vary in a consistent fashion, and widely available it was difficult to fit the VBGF to growth and maturation parameters are closely length-at-age data, and several methods were de- interrelated (Adams 1980; Charnov and Berrigan veloped for the estimation of L• and k (Fig. 5.4). 1991; Beverton 1992; Gunderson 1997; He and One method involves making a plot of the annual

Stewart 2001; Hutchings, Chapter 7, this volume). increment of length (Lt+1 - Lt) against length (Lt), As such, it may be possible to estimate both the age where Lt+1 is length at age t + 1 and Lt is length at and size at first reproduction from the information age t. This gives a straight line with a slope of encapsulated in a growth trajectory plot (Charnov -(1 - e-k), and an intercept on the abscissa (i.e. and Berrigan 1991; He and Stewart 2001). Charnov where Lt+1 - Lt = 0) equal to L•. This equation is also and Berrigan (1991) and He and Stewart (2001) known as the Brody equation, as mentioned by summarized the interrelationships and provided Schnute and Richards (Chapter 6, Volume 2). This models, Beverton (1992) considered the relation- expression not only establishes the constants in ships within the framework of the growth– the VBGF but also provides an indication of the de- maturity–longevity (GML) plot, and Gunderson cline in the rate of growth with age. The constants

(1997) reviewed the relationships for viviparous can also be estimated from a plot of Lt+1 on Lt, the and oviparous fish species: viviparity seemed to be Ford–Walford plot (Fig. 5.4). The rate of growth of associated with reduced reproductive effort, in- the fish slows with age so the plotted line gradually creased age at maturity and low mortality relative approaches a 45° line passing through the origin. to oviparous species of similar size. The two lines will intersect at L•, the point of in- The von Bertalanffy growth function (VBGF) tersection indicating when the lengths of the fish

fits many observations on the growth of fish and is at the start (Lt) and end (Lt+1) of the growth period widely used in fisheries research because the con- are identical, i.e. the fish has ceased to increase in stants were readily incorporated into early stock length, and the annual growth increment is zero. assessment models (Beverton and Holt 1957). The The growth constant, k, can also be estimated from simplest derivation of the VBGF is: the plot of Lt+1 on Lt because the slope of the line is equal to e-k. Much work has been done on develop- ddLt=- kLL()• (5.4) ing methods for fitting and testing VBGF data, and with the advent of the personal computer the where L is length, t is time, L• is the asymptotic handling of the data has become much easier length, which is the length the fish would reach if (Gallucci and Quinn 1979; Misra 1986; Ratkowsky it were to grow to an infinite age, and k is the 1986; Cerrato 1990, 1991; Xiao 1994). For a gener- growth constant expressing the rate at which alization of the von Bertalanffy model see Schnute length approaches the asymptote (Gulland 1969; (1981).

Ricker 1979). Whereas L• has a straightforward Bayley (1977) pointed out that a weakness in interpretation, that of k is less easy because it several of the methods is a lack of independence describes the instantaneous growth rate (dL/dt) between the variables plotted. In an attempt to relative to the difference between L• and the overcome the problem, Bayley (1977) devised a length of the fish at a given time. Integration of the method for the estimation of the VBGF constants

VBGF gives: (L• and k) using measurements of instantaneous growth rates (d(lnW)/dt) and a description of LL=-1 e--kt() t0 (5.5) the length–weight relationship (W = cLm). From t •() 104 Chapter 5

(a) 10 y = 10.422 – 0.1001x or R 2 = 0.3232 8 ()lnW2121- ln W() t- t=- mk + mkL•()1 L (5.7) 6 The latter has the form of a linear regression with a

slope of mkL•, the intercept is -mk, and a plotted 4 line will intersect with the abscissa at 1/L•, where, by definition, the instantaneous growth rate is 2 zero. Thus, the constants of the VBGF can be estimated from successive measurements of 0 0 20 40 60 80 100 120 length and weight, and calculation of m in the Annual length increment (cm) length–weight relationship; instantaneous growth –2 Length (cm) at age t (years) rate is plotted against the reciprocal of fish length, –4 the slope and intercept of the regression calcu- lated, and the VBGF constants are then estimated from the values obtained. Bayley (1977) suggested (b) 120 that this method of analysis could be appropriate y = 0.8999x + 10.422 for the estimation of the VBGF constants for tropi- 100 R 2 = 0.9748 cal fish species in which age determination may be 80 extremely difficult. For these species growth is + 1 (years)

t often estimated from data collected following the 60 recapture of released marked fish, where it is not usually possible to control the time over which in- 40 dividual fish are at liberty. Analysis of growth data using this method does, however, require that 20 there has been a marked change in fish weight and Length (cm) at age length over the growth period. 0 0 20 40 60 80 100 120 Length (cm) at age t (years) 5.5 AGE DETERMINATION, Fig. 5.4 Graphical methods for estimating the BACK-CALCULATION AND parameters L• and k of the von Bertalanffy growth function using data for female spiny dogfish, Squalus VALIDATION TECHNIQUES acanthias, as an example (see Fig. 5.3 for length at age plot). (a) A plot of the annual length increment against The ability to determine the age of fish is an impor- length gives a straight line with a slope of -(1 - e-k) and tant tool in fisheries research, and age data are vital an intercept on the abscissa at L•. (b) A Ford–Walford for both growth modelling and the study of popula- -k plot gives a regression line (solid line) with a slope of e . tion dynamics. There are several approaches to The point of intersection of the lines indicates L . • ageing fish, including ‘direct’ observation of indi- viduals, length–frequency analysis and the analy- sis of various hard structures (Bagenal and Tesch 1978; Weatherley and Gill 1987; Busacker et al. these data Bayley (1977) derived an equation that 1990; Devries and Frie 1996; Campana 2001). led to a linear transformation of the non-linear The most accurate method for collecting age VBGF: data is direct observation of individuals, but this is time-consuming and costly. However, under some dd()lnWtmLkLL= ()[]()••- = mkLL[]()-1 circumstances it may be the only way in which re- (5.6) liable age and growth information can be obtained. Rates of Development and Growth 105

The method involves the release of marked fish ticularly valuable in validation studies designed into natural systems. Marked fish may be either to cross-check fish ages as determined by other hatchery-reared fish of known age or fish captured methods (Weatherley and Gill 1987; Brown and and marked in situ. The marked fish are later re- Gruber 1988; Casselman 1990; Rijnsdorp et al. captured. The period of time between release and 1990; Devries and Frie 1996; reviewed by Campana recapture is quantified and combined with data re- 2001). lating to changes in body size for use in growth Length–frequency analysis may be used to dis- models. Data collected using the mark-and- tinguish between different age groups of fish pro- recapture method can also be used to gain insights vided that the distribution of lengths is unimodal into the size of the fish population, provided that around a value that is distinct for the age group. certain assumptions are met (Youngs and Robson Several graphical and statistical methods have 1978; Guy et al. 1996). One prerequisite of the been developed to enable age groups to be sepa- method is that the released fish can be easily recog- rated using an analysis of length–frequency data nized at the time of recapture. In other words, the (reviewed by MacDonald 1987; Pitcher, Chapter 9, mark applied must be distinct and, if not perma- Volume 2). Although length–frequency analysis nent, at least long-lasting. Further, if data from appears simple in principle, it may not be so in marked fish are to be of value in the estimation of practice. Slower and disparate growth amongst growth rates and population sizes, the marks ap- older fish makes for less distinct peaks at older plied should not influence either growth or the vul- ages, and long spawning seasons lead to an increase nerablity of the fish to predators, i.e. mortality in the spread of the distributions of length (Bagenal rates should not be affected. and Tesch 1978; Devries and Frie 1996). Several marking and tagging techniques are Although direct observation and length– available (Laird and Stott 1978; Guy et al. 1996; frequency analysis are used in age determination Campana 2001). Fish may be marked by fin mutila- studies, the most frequently used method is the ex- tion, hot and cold branding or tattooing; marks amination of hards parts such as scales, otoliths, may also be applied to the fish via subcutaneous in- spines, vertebrae and the opercular and dentary jections of dyes, liquid latex or fluorescent materi- bones (Bagenal and Tesch 1978; Weatherley and als. There are also many types of tags, some of Gill 1987; Brown and Gruber 1988; Casselman which, such as the anchor tag and the plastic flag 1990; Rijnsdorp et al. 1990; Devries and Frie 1996; tag, are applied externally, whereas others, such as Campana 2001). This method is based on the fact visible implant tags, coded wire tags and passive that as a fish grows by increasing in length there integrated transponders (PIT tags), are subcuta- will also be increases in the size of the hard body neous or internal. The advantage of tags over other parts (Table 5.1). However, the use of hard parts for marking techniques is that tags can be numbered age determination also relies on the appearance of serially, allowing for individual recognition. growth zones, rings or checks in these parts. These Chemical marks may be induced in the body tis- growth marks are referred to as annual marks, sues by feeding, injecting or immersing the fish in annual rings or annuli when they are taken to solutions of a chemical that is taken up and incor- represent age in years, and daily rings or daily porated into the tissue in question. The hard calci- increments when they indicate daily growth. fied tissues, such as scales, otoliths and skeletal For example, the number of scales of a bony fish elements, are the most common tissues used be- remains nearly constant throughout life, so as the cause they incorporate certain chemicals perma- fish grows the scales must inevitably increase in nently and in a form that can provide a ‘time mark’. size in more or less the same proportion. New ma- Examples of chemical markers include fluorescent terial is added to the scale as it increases in size: in compounds such as tetracycline and calcein, and the elasmoid scales (cycloid or ctenoid) of teleost metallic elements such as strontium and rare earth fishes this material appears as a series of bony elements. Chemical marking techniques are par- ridges, or circuli. The arrangement of the circuli is 106 Chapter 5

Table 5.1 Relationships between otolith length (OL) and fish standard length (FSL) (both in mm) for some common boreal marine fish species. (Source: from Jobling and Breiby 1986.)

Fish species n FSL size range (mm) FSL = a + b OL

Intercept (a) Slope (b)

Herring, Clupea harengus 76 51–279 -8.50 58.46 Pearl-side, Maurolicus muelleri 74 22–67 -12.80 35.31 Lanternfish, Benthosema glaciale 37 37–85 -3.31 31.79 Capelin, Mallotus villosus 96 49–159 24.02 44.31 Sandeel (sand-lance), Ammodytes tobianus 42 65–137 14.93 50.81 Atlantic cod, Gadus morhua 67 26–317 0.41 22.44 Haddock, Melanogrammus aeglefinus 65 20–283 11.06 17.24 Saithe (coal-fish), Pollachius virens 37 53–259 -4.24 23.50 Blue whiting, Micromesistius poutassou 27 56–318 -25.54 23.17 Norway pout, Trisopterus esmarki 75 72–252 -23.26 25.14 Silvery pout, Gadiculus argenteus thori 37 79–149 4.53 18.14 Red-fish (Norway haddock), Sebastes spp. 7818–238 3.75 19.45

not regular because the fish does not grow at the sampling (Casselman 1990; Devries and Frie 1996; same rate throughout the year. When food is plen- Campana 2001). tiful, the fish grow rapidly and the scales increase It may be difficult to determine the age of tropi- in size by the addition of large numbers of circuli cal and subtropical fish species by examination of well separated from each other; when growth the scales or other hard parts, because the growth slows down or ceases the circuli are much closer ‘rings’ may not necessarily be annual. The for- together. Thus, periods of slow growth are denoted mation of ‘rings’ may be associated with factors by a close clumping of circuli to form a distinct unrelated to specific seasonal changes, such as ‘ring’. In temperate regions, the growth of fish is unpredictable changes in food supply, spawning usually rapid during the spring and summer but is events or density-related changes in growth slow during the winter. These seasonal changes (Bagenal and Tesch 1978; Weatherley and Gill are reflected in the deposition of circuli in the 1987; Devries and Frie 1996). Thus, if age determi- scales. Thus, the winter growth check enables nation is attempted using readings of these growth the age of the fish to be determined by counting the ‘rings’ there must be some sort of validation by number of ‘rings’ formed by areas of close circuli other methods (reviewed by Campana 2001). on the scales. The greater the seasonal growth dif- Unfortunately it may not be possible to use ferences, the clearer the growth ‘rings’ on the scale: length–frequency analysis because the fish may for this reason they are most easily seen on the spawn throughout much of the year and life cycles scales of fish from high latitudes, where growth is are short, so a ‘direct’ method using mark-and- distinctly seasonal. Thus, scales may be the hard recapture may be the method of choice. part of choice for ageing fish from high latitudes, Similarly, it is not possible to determine the age and they have the added advantage that their col- of elasmobranchs using scales because, unlike the lection need not involve killing the fish. However, scales of bony fishes, the placoid scales of sharks scales can be demineralized during prolonged peri- and rays do not increase in size as the fish grows. ods of food deprivation and scales can be regener- Instead, new scales are added in between the exist- ated after damage; this can lead to errors in age ing ones. This is a continuous process so that at any determination unless sufficient care is taken with one time there will be a number of newly erupted Rates of Development and Growth 107 scales, as well as some in the process of disintegra- is the radius of the hard part at time t, and a is the tion. However, age determination of some elasmo- intercept of the regression line relating hard-part branchs may be possible by examination of growth radius to fish length (Bagenal and Tesch 1978; zones within vertebrae or spines (Holden and Devries and Frie 1996). In some cases, the use of Meadows 1962; Holden 1974; Nammack et al. simple linear regression may be precluded due to a 1985; Brown and Gruber 1988). lack of linearity between the dimensions of the When the scales and other hard parts increase in hard parts and the body, or because there are differ- size in proportion to the size of the fish (e.g. Table ent body length to hard-part relations among age 5.1), they may not only be used in age determina- groups. Under such circumstances various curve- tion but can also be considered to represent a diary fitting procedures and covariance analysis may be recording the growth history of the fish. Thus, used to address the problems (Bagenal and Tesch using knowledge about the relationship between 1978; Bartlett et al. 1984). The Weisberg method is the size of the hard part and fish length, it may be more complex than the others. It involves a model- possible to back-calculate the length of the fish at a ling approach that enables age group and annual given age by examination of the positioning of the environmental effects to be distinguished. Thus various growth ‘rings’ (Bagenal and Tesch 1978; there is a separation and identification of changes Weatherley and Gill 1987; Devries and Frie 1996). in growth from one time period to another, such as The data required for back-calculation are the years of particularly good or poor growth, that may length of the fish at capture, the radius of the hard be superimposed upon age effects (Weisberg and part at capture (measured from the nucleus to the Frie 1987; Weisberg 1993). margin), and the radius of the hard part to the outer edge of each of the growth ‘rings’ (either annuli or daily increment rings). Back-calculation of length 5.6 LENGTH–WEIGHT at any given age is then usually carried out by one RELATIONSHIPS, AND of four methods: the direct proportion method, the INDICES OF CONDITION Fraser–Lee method, various curve-fitting proce- AND GROWTH dures or the Weisberg method (Bagenal and Tesch 1978; Devries and Frie 1996). The method to use The length of a fish can be measured more easily depends upon the type of relationship between the and accurately than weight under field conditions, length of the fish and the dimensions of the hard so length measurements are the most convenient part used in the back-calculation procedure. method of growth expression. Frequently, infor- The direct proportion method can be used when mation about fish weight may be required and this the relationship between body length and hard- may be estimated from length if the length–weight part radius is linear and has an intercept that does relationship is known for the fish population not differ from the origin: in this situation the under study. The length–weight relationship is growth of the hard part is directly proportional to generally expressed by the equation: the growth in length of the fish. The Fraser–Lee method is applicable when the intercept of the m relationship between fish length and hard- W = cL (5.9) part radius is not at the origin. Under these or circumstances, length at time t (Lt) can be back- calculated using the formula: logW = logc + mlogL (5.10) LLaRRatc=-[]() ct+ (5.8) where W is weight, L is length, and c and m are con- where Lc is the length of the fish at the time of cap- stants. The numerical value of m is nearly always ture, Rc is the radius of the hard part at capture, Rt between 2.5 and 3.5, and is often close to 3 (Bagenal 108 Chapter 5 and Tesch 1978; Weatherley and Gill 1987; relationship is close to 3, the condition index is Anderson and Neumann 1996; Sutton et al. 2000). usually calculated as Fulton’s condition factor (K): When m = 3, the body is increasing in all dimen- sions in the same proportions as it grows. How- KIWIL= 3 ¥ 100 (5.12) [] ever, the relationships of the different body dimensions often change with respect to each In most cases K will be satisfactory for analysing other as the fish grows, so body shape changes as differences in condition related to sex or season the fish increases in length. This means that m will when the fish used for comparison are of approxi- be greater than 3 when a fish becomes more rotund mately the same length; however, if the length as it increases in length, and less than 3 if the fish range is large, spurious results will be generated if becomes ‘slimmer’ as it increases in body length. m differs from 3. As a fish grows, changes in weight are relatively Changes in growth rate and condition are ac- greater than changes in length, due to the approxi- companied by changes in the biochemical com- mately cubic relationship between length and position of the body tissues; the most marked weight. Thus, measurement of change in weight changes occur in percentages of lipids and body may provide a more precise measure of growth moisture, whereas the relative proportion of pro- over short periods of time. However, a change in tein tends to vary to a lesser extent (Weatherley weight may be a very transient indicator of growth, and Gill 1987; Love 1988; Hislop et al. 1991; because weight can be both gained and lost. Thus, Shearer 1994; Brett 1995; Couture et al. 1998; when fish are feeding and growing well, an individ- Dutil et al. 1998; Sogard and Olla 2000; Sutton ual may have a greater than usual weight at a par- et al. 2000). As a consequence of this, there is a ticular length. On the other hand, when feeding strong negative correlation between the percent- conditions are poor, the fish may lose weight and ages of body lipids and body moisture (Fig. 5.5), and be light for their length. Thus although length is this relationship seems to hold under different the primary determinant of the weight of a fish, conditions of feeding, growth and gonad develop- there can be wide variations in weight between ment (Weatherley and Gill 1987; Love 1988; fish of the same length both within and between Hislop et al. 1991; Sogard and Olla 2000; Sæther populations. Thus, length–weight relationships and Jobling 2001). Thus, temporal changes in the can be used to assess the ‘well-being’ of individual condition of the fish, related to the deposition and fish. Since it is not particularly easy to interpret mobilization of energy reserves, will be reflected and directly compare the constants in the by movements up and down the lipid–moisture (or length–weight relationship, a series of indices of fat–water) regression line. condition have been developed in an attempt to Attempts have been made to use a range of circumvent some of the problems. metabolic variables to obtain an assessment of The length–weight relationship given by the condition and recent growth history of wild equation 5.9 can be rearranged to give an index of fish (Busacker et al. 1990; Houlihan et al. 1993; condition, or condition factor (CF): Couture et al. 1998; Dutil et al. 1998). The meta- bolic indicators monitored include concentrations CF= IW ILm ¥ 100 (5.11) of nucleic acids, and enzymatic indicators of aero- [] bic and glycolytic capacities of muscle, liver and where IW and IL are the weight and length of an in- intestine. Tissue RNA:DNA ratio has often been dividual fish respectively, and m is the exponent in used both to assess fish condition and as an indi- the length–weight relationship. Alternatively, the rect measure of recent growth. It has been argued condition of an individual fish can be calculated as that because the DNA content of a cell is relatively IW/EW, where EW is the ‘expected weight’ of the constant, and RNA content varies with the rate of fish calculated from the length–weight relation- protein synthesis, the ratio of RNA to DNA pro- ship. In practice, because m in the length–weight vides an index of protein synthetic activity and Rates of Development and Growth 109

78 biochemical indices, such as protein synthesis, tis- sue amino acid incorporation and RNA:DNA 76 ratios, used for the assessment of condition and short-term changes in fish growth. Intestinal mi- 74 tochondrial enzyme activity also seems to corre- late with rates of feeding and growth (Couture et al. 1998; Dutil et al. 1998). Nutrient absorption by the 72 enterocytes of the gut involves active transport; this relates to Na+/K+ATPase activity, and 70 the enterocytes are mitochondria-rich cells in

Percentage water (moisture) which ATP is produced aerobically (Ferraris and 68 Diamond 1989; Hirst 1993). Consequently, links between intestinal mitochondrial enzyme activ- 66 ity, nutrient absorptive capacity and growth might 2 4 6 810 12 14 be expected. Percentage fat (lipid) Prolonged fasting affects white glycolytic mus- cle more than oxidative muscle, and when food Fig. 5.5 The percentage of body fat (lipid) and water deprivation lasts a period of weeks there are often (moisture) in samples of turbot, Scophthalmus marked declines in muscle sarcoplasmic proteins maximus, at the start of a growth trial (; weight c.200 g), after feeding for 12 weeks (weight c.450g) on either a and glycolytic enzyme activities (Love 1988; high-fat (, 25% lipid) or low-fat (, 17% lipid) diet, and Couture et al. 1998; Dutil et al. 1998). Conversely, then after 8 weeks (weight 550g) of dietary reversal (i.e. during periods of intense feeding the activity of , high-fat to low-fat diet; , low-fat to high-fat diet). white muscle glycolytic enzymes, such as phos- The ‘fat–water’ line is described by the following regres- phofructokinase, pyruvate kinase and lactate de- 2 sion: %water = 80.11 - 1.08 %fat (n = 44; R = 0.866). hydrogenase, may be positively correlated with (Source: data from Sæther and Jobling 2001.) feeding and growth rates. Activities of mitochon- drial enzymes, such as cytochrome c oxidase and hence growth (Bulow 1987). In addition to citrate synthetase, in the oxidative red muscle ap- RNA:DNA ratio, tissue RNA content and concen- pear to be poor predictors of growth. This is not sur- tration and RNA:protein ratio have been used as prising given the fact that during fasting it is the indirect indices of fish condition and growth. In white muscle proteins that are mobilized, whereas studies with fish larvae it is usually the RNA: the aerobic red muscle seems to be preferentially DNA ratio that has been used as the growth corre- conserved (Love 1988). The high activities of gly- late, and results of a study on cod larvae provide ev- colytic enzymes in the muscle and their sensit- idence that the analysis of nucleic acids may ivity to food supply seem to make them useful provide valuable information about the recent indicators of condition and growth, with the moni- growth and condition of individual larvae (McNa- toring of changes in lactate dehydrogenase activity mara et al. 1999). appearing to be particularly useful (Couture et al. Another biochemical component with the po- 1998; Dutil et al. 1998). tential to correlate with recent growth history is ornithine decarboxylase activity (Benfey 1992; 5.7 ENERGY BUDGET Benfey et al. 1994). Ornithine decarboxylase is the first, and rate-limiting, enzyme in the biosynthesis AND BIOENERGETICS: of polyamines, compounds essential for the ENERGY PARTITIONING biosynthesis of nucleic acids and proteins. Thus, it AND STORAGE might be expected that changes in ornithine decar- boxylase activity would precede changes in other In simple terms, growth is the change that results 110 Chapter 5 from the difference between the food that enters The proportion of the food energy lost in the faeces the body and the waste materials that leave it and nitrogenous excretory products may also be (Brett 1979, 1995; Cho et al. 1982; Weatherley and influenced to some degree by the amount con- Gill 1987; Adams and Breck 1990; Elliott 1994; sumed, fish size and temperature (Elliott 1994; Jobling 1994, 1997). This can be represented by: Jobling 1994). The metabolic term, M, encompasses fasting p C = M + G (5.13) metabolism, activity metabolism and feeding me- tabolism. The latter is the metabolism linked to or the processing of nutrients and the elaboration of tissues. Metabolic rates are usually assessed by an G = p C - M (5.14) indirect method involving the measurement of oxygen consumption and the application of an where C is the amount of food consumed, p is a co- oxycalorific coefficient (1ml O2 = 19.4J) (Adams efficient indicating the availability of nutrients or and Breck 1990; Cech 1990; Jobling 1994; Brett food energy, M represents catabolic losses (metab- 1995). In an expanded model each metabolic com- olism) and G is the anabolic component, the nutri- ponent will be described using a separate mathe- ents or food energy retained as growth. Implicit in matical function, which will include an allometric this energy budget equation is the dependency of term to account for size effects and will also con- growth on consumption, so the intake of food tain a temperature function: energy is the pacemaker of growth. However, the increase in growth with increased food intake may M = aWbecTedS + eC (5.16) not be monotonic (Brett 1979; Weatherley and Gill 1987; Elliott 1994; Jobling 1994, 1997). where W is body weight, S is swimming speed and Temperature is the most all-pervasive environ- a–e are constants. The weight exponent, b, for mental factor that influences aquatic organisms fish is about 0.8, and the temperature constant, c, and, via its influences on feeding and metabolism, will generally lie within the range 0.04–0.07, cor- it affects growth. Increases in temperature initially responding to a Q10 of 1.5–2.0. Metabolic costs lead to an increase in food consumption; feeding associated with activity show an approximately peaks at some intermediate temperature, and then exponential increase, whereas the metabolic costs declines precipitously as the temperature contin- associated with food processing, tissue synthesis ues to rise (Fig. 5.6). It may be difficult to describe and energy storage are proportional to the quantity the influence of temperature on food intake in of food consumed. mathematical terms, but Hogendoorn et al. (1983) The difference between the rate–temperature suggested that Hoerls function could be used to de- curves for feeding and metabolism gives an indica- scribe the effects of temperature on food intake of tion of the resources available for growth under fish provided with unrestricted rations: different temperature conditions, assuming that p represents a constant proportion of C across C = aTbecT (5.15) temperatures. The plotting of the growth rate– temperature relationship indicates that growth where T is temperature and a, b and c are rate reaches a peak at an intermediate temperature constants. (Fig. 5.6), with the optimum temperature for The proportion of the ingested energy lost as growth being slightly lower than that at which faecal and excretory wastes depends upon the na- food consumption rate reaches its maximum. Re- ture of the diet (Cho et al. 1982; Jobling 1994), so lationships of this type have been obtained in a the value of p differs between fish species that number of growth studies conducted on fish (e.g. differ in trophic habits. For carnivorous species a Brett 1979; Hogendoorn et al. 1983; Woiwode and value of p close to 0.8 may be appropriate, but for Adelman 1991; Xiao-Jun and Ruyung 1992; Elliott herbivores and detritivores values will be lower. 1994; Liu et al. 1998; reviewed by Jobling 1997), but Rates of Development and Growth 111

high-latitude environments, and there may also be changes in composition that are directly related to body size (Weatherley and Gill 1987; Love 1988; Hislop et al. 1991; Shearer 1994; Jørgensen et al. 1997; Van Pelt et al. 1997; Sogard and Olla 2000; Sutton et al. 2000). In other words, fish tend to ac- cumulate storage lipids during the summer growth season, and the reserves are then mobilized to provide metabolic fuel during the winter, when (A) food consumption is low, and to support gonadal Metabolic rate ( ) Ingestion rate ( ) development (Brett 1979, 1995; Weatherley and Gill 1987; Love 1988; Elliott 1994; Jobling 1994; Jørgensen et al. 1997). Upper The feeding and growth responses seen in high- thermal latitude fish during the summer months resemble limit for those observed amongst fish that have been de- growth (B) prived of food under laboratory conditions: in-

Growth rate creased food intake, or hyperphagia, rapid growth and improvement in condition, and the repletion of energy reserves (Weatherley and Gill 1987; Temperature Broekhuizen et al. 1994; Jobling 1994; Jobling and Fig. 5.6 Rate–temperature curves illustrating the Johansen 1999; Sæther and Jobling 1999). It is usu- effects of temperature on rates of ingestion, metabolism ally fish that are in poor condition that show and growth. Note that the temperature at which ingestion rate reaches its maximum (A) is a few degrees the greatest response (Fig. 5.7)(Jobling et al. 1994). higher than the optimum temperature for growth (B). Thus, seasonal cycling may relate to a regulation of (Source: from Jobling 1997.) the balance between reserves held in the form of lipid depots and mobilizable parts of the muscula- ture and structural components such as the skele- ton and circulatory and nervous tissue. Shifts in some authors have reported that the temperatures the balance would be reflected in relative changes at which feed intake and growth peak are almost in major chemical components that constitute the coincident (e.g. Larsson and Berglund 1998; tissues, for example in the relative proportions Forseth et al. 2001) of body lipid and moisture (cf. Fig. 5.5). Large The growth term, G, encompasses somatic shifts would be expected to induce compensatory growth related to tissue elaboration, the storage changes in feeding and energy accumulation. Fol- and mobilization of energy reserves, and reproduc- lowing restoration of the balance between com- tive growth; the latter may be difficult to assess ac- partments, rates of feeding would be predicted to curately for wild fish. Thus, the changes relating to decrease and rates of growth and energy accumula- growth over time include the energy deposited as tion would slow (Broekhuizen et al. 1994; Jobling protein and lipid in the soma, the changes in the and Johansen 1999). This presupposes that specific size of the energy storage depots and the energy di- metabolic signals arising from different tissues are rected towards the production of gametes. Protein integrated within the central nervous system, en- growth may be positive throughout much of the abling continuous assessment of the status of the year but quantities of lipids, and gonad sizes, may energy reserves (reviewed by Kiess et al. 1999; undergo marked fluctuations, including large Magni et al. 2000). net losses at certain times of the year. Seasonal In studies of fish growth the energy content changes in body composition will usually be most of the tissues is usually determined by bomb extreme in fish that inhabit temperate-zone or calorimetry. Bomb calorimetry involves the rapid 112 Chapter 5

18 weeks combustion of the sample in oxygen at increased 6 pressure, and heat production is measured. This provides a measure of the heat of combustion, or gross energy, of the sample. Alternatively, the 4 chemical composition of the sample can be meas- ured using a series of standard laboratory methods (Osborne and Voogt 1978; AOAC 1990; Busacker Weight (kg) et al. 1990), and conversion factors for the com- 2 plete oxidation of proteins, lipids and carbohy- drates used to estimate gross energy content. 6 weeks Complete oxidation of proteins, lipids and carbo- hydrates yields approximately 24, 39 and 17kJg-1 4 respectively, although the energy yield from fish lipids may be slightly lower than that from lipids of 3 terrestrial origin. The chemical analysis of the major components (moisture, protein, carbohy- 2 drates, lipids and ash) of animal tissues is usually

Weight (kg) termed a proximate analysis. 1 A characteristic of the growth of many fish species in nature is the marked seasonal variabil- 0 ity. This seems to be almost universal outside of Start the tropics and is by no means rare within them; in the tropics and subtropics the variations in fish growth may be most closely related to seasonal 2 changes in rainfall. It is those species that live at high latitudes that exhibit the greatest seasonal fluctuations in feeding and growth. These seasonal

Weight (kg) 1 growth cycles are often attributed to the temporal changes in food availability that may occur in tem- perate and polar waters, but even in the absence of 0 changes in food availability low water tempera- Group 123456tures during the winter months would be expected Condition <0.80.8–0.9 >0.9 <0.80.8–0.9 >0.9 to restrict growth (Fig. 5.6). However, the seasonal (start) cycle of growth is not always, and perhaps never, Length (cm) <60 60–70 wholly under temperature control; the growth of (start) several high-latitude fish species seems to track Fig. 5.7 The effects of initial length (L cm) and the seasonal cycle of photoperiod (Brett 1979; condition (K = [W/L3] ¥ 100) on weight (W) change in Woiwode and Adelman 1991; Jobling 1994; Boeuf groups of Atlantic cod, Gadus morhua, during 18 weeks and Le Bail 1999). This comes particularly to the of growth. Note that the cod that were in the poorest fore when the fish are exposed to constant temper- condition at the start displayed the highest rates of ature and unrestricted food supply. It is possible weight gain, and that at 18 weeks the body weights that photoperiod acts to synchronize an endoge- attained by cod within each initial length group were similar. The boxes indicate 50% of the ‘population’ and nous growth rhythm with prevailing environmen- the bars the 95% confidence limits. (Source: from Jobling tal conditions. In this case photoperiod acts as a et al. 1994.) synchronizing timing signal, or zeitgeber. The re- sponse to photoperiod would be expected to be marked amongst high-latitude species, because it Rates of Development and Growth 113 is at these latitudes that photoperiodic signals are potential. Thirdly, there may be compensatory strongest and aquatic habitats are most variable on mechanisms that operate to counteract the a seasonal basis. However, because the natural sea- negative effects of low temperature and a short sonal variations in food availability, temperature growth season on the growth of fish that inhabit and daylength tend to follow similar cycles, it may high-latitude environments. not be easy to distinguish the effects of each vari- There are two basic models of how compensa- able on growth and energy partitioning. tory mechanisms might be expressed. One model relates the compensatory mechanisms to local thermal adaptation, such that growth rates are 5.8 GROWTH AT maximized at temperatures commonly experi- enced by fish within their native environment. DIFFERENT LATITUDES: According to this model there would be differences MODELS OF GROWTH in optimum temperatures for growth amongst COMPENSATION populations inhabiting different thermal environ- ments, and the compensation should be accom- Many species of fish have wide geographic distri- panied by a change in preferred temperatures (Fig. butions; populations of these species occur at 5.8) (Jobling 1997). The second model focuses on different latitudes and inhabit environments that latitudinal differences in seasonality rather than differ in both mean annual temperature and length temperature per se (Conover 1990, 1992; Conover of the growing season. For many of these species and Schultz 1995). In this model, high-latitude fish the annual growth increment is lower for fish have a higher capacity for growth to compensate from higher latitudes than for those living at low for short growing seasons: latitudinal compensa- latitudes, but the differences appear to be less tion is observed as an elevation in the growth pronounced than expected from the differences in rate–temperature curve (Fig. 5.8). According to temperature and growth conditions experienced this model, genetic and environmental influences by the different populations (Conover 1990, 1992). oppose one another along the gradient of decreas- For example, Conover (1990, 1992) provided ing length of the growing season, i.e. the fastest- evidence that there were minimal differences in growing genotypes are found in environments size at the end of the first growing season for that have the most depressive effect on growth several species of fish that occur on the eastern (Conover and Schultz 1995). seaboard of North America: Atlantic silverside There are two ways in which these models (Menidia menidia), American shad (Alosa sapidis- could be tested. One method involves common- sima), striped bass (Morone saxatilis) and mummi- garden experiments in which fish from different chog (Fundulus heteroclitus). There are three ways populations are reared under the same tempera- in which such a situation could arise, which are ture range, and growth rate–temperature curves not mutually exclusive. Firstly, size-selective are plotted for each test population (e.g. Jonassen mortality could be more pronounced in high- et al. 2000). The second method involves recipro- latitude populations leading to survival of only cal transplants, i.e. representatives from the differ- the largest fastest-growing individuals, whereas ent test populations are reared in each of the a greater proportion of the slower-growing fish original habitats in order to examine the (pheno- might survive at lower latitudes: the net result typic) responses of the different populations would be little or no difference in mean size of the (prospectively different genotypes) to a combina- fish at the end of the growth season. Secondly, tion of environmental influences. growth of fish in the populations at low latitude The results of transplantation experiments might be subject to greater constraint by food have been equivocal, but when reared under the availability, so that fish in these populations grow same conditions fish from high-latitude environ- at rates substantially below their full physiological ments may grow faster than conspecifics from 114 Chapter 5

(a) both a higher growth capacity and a lower opti- mum temperature for growth than did halibut from lower-latitude populations. Additional evi- dence that there may be an elevation of the growth rate–temperature curve in fish from high-latitude populations has been obtained in studies conduct- ed over more limited temperature ranges (Schultz et al. 1996; Conover et al. 1997; Brown et al. 1998; Imsland et al. 2001). Thus, growth compensation (b) Growth rate across the geographic range of a species may not be based simply on genotype–environmental temper- ature interactions, and it has been hypothesized that latitudinal growth compensations may have evolved to offset the disadvantages of being small at the end of the growing season (Conover 1990, 1992). Although the ability to grow rapidly within a short growing season would clearly be advanta- geous for fish in high-latitude environments, there Temperature would also need to be forces operating in a negative Fig. 5.8 Rate–temperature curves illustrating two ways direction to enable latitudinal variation in growth- in which compensatory adjustments could be made to rate capacity to be maintained (Conover and counteract the effects of environmental differences on Schultz 1995). Such forces might, for example, re- growth. (a) The rate–temperature curves are displaced late to increased incidences of developmental de- in relation to each other; fish from low-temperature formities or to increased metabolic costs, or could environments have a lower optimum temperature for encompass trade-offs involving negative cor- growth, and grow faster at low temperatures, than do fish from warm-water populations (thermal adaptation). relations between growth capacity and disease (b) Compensation is envisaged to occur by means of resistance, or the ability to withstand prolonged an elevation of the rate–temperature curve; fish with periods of food shortage. Differences in metabolic the highest growth capacity are found in environments rate related to latitude of occurrence have often that have the greatest inhibitory effects on growth been reported, and the metabolic compensations (countergradient variation). Bars indicate the zones have usually been interpreted as being a response of thermal preferenda predicted for fish from different to temperature (Cossins and Bowler 1987). The po- environments. (Source: from Jobling 1997.) tential benefits of metabolic compensation in rela- tion to temperature are not immediately clear, but lower latitudes. This would tend to call into ques- there are links between metabolic rate and the tion the idea that there is a marked downward ability to maintain activity and growth (for dis- shift in the optimum temperature for growth in cussion see Hochachka 1988; Clarke 1991, 1993, high-latitude populations (Fig. 5.8). Results from 1998). Consequently, an elevated metabolic rate some common-garden experiments designed to may be a prerequisite for exploitation of food test growth rate–temperature responses add resources over a short growth season. Thus, the support to this: optimum temperatures for growth possibility arises that differential metabolic costs of different populations have been found to be might represent a trade-off leading to the mainte- similar (McCormick and Wegner 1981; Conover nance of latitudinal variation in growth-rate and Present 1990). However, Jonassen et al. (2000) capacity. Interactions of this sort would have con- reported that juvenile halibut (Hippoglossus hip- sequences for bioenergetic modelling, because it is poglossus) from a high-latitude population had usually assumed that such models can be con- Rates of Development and Growth 115 structed using a single set of species-specific phys- tion that has fed on a particular prey type. This pro- iological parameters. In other words, the type of vides an indication of the uniformity with which trade-off described above invalidates the assump- fish within a population select their diet, but it tion of invariant species-specific parameters, and does not give any information about the relative an accurate modelling of energy flow and nutrient importance of the different prey types. The calcu- partitioning would require population-specific lation of percentage composition by number gives knowledge to be available. the latter sort of information because it provides an estimate of the relative abundance of a particu- lar food item in the diet. Percentage composition 5.9 ESTIMATING FOOD by number has its equivalents in percentage com- CONSUMPTION position by volume and percentage composition by weight, obtained from the results of volumetric Estimation of food consumption by fish popula- and gravimetric analyses respectively (Windell tions has widespread uses in ecological and fish- and Bowen 1978; Bowen 1996). There are weak- eries research. Such estimations are required for nesses associated with all these forms of analysis, the investigation of predator–prey interactions and in an attempt to overcome the problems there and predation mortality, for assessment of produc- have been several attempts to develop indices that tion dynamics of fish populations, and for the combine the information obtained using the dif- development of multispecies stock assessment ferent analyses (Bowen 1996). models (Ney 1990; see also Shepherd and Pope, In many cases it may be deemed desirable to Chapter 7 and Pauly and Christensen, Chapter 10, have estimates of the size, or weight, of the differ- both Volume 2). For accurate estimation of popula- ent prey items at the time of ingestion. Sizes of par- tion food consumption many different types of tially digested prey can be estimated by measuring data are required. These include information the dimensions of some digestion-resistant hard about population density and age structure, as well part such as an otolith, operculum, head capsule, as quantitative assessments of food consumption carapace or chitinous element of a compound rates of fish of different sizes under different envi- appendage. The initial size of the prey can then be ronmental conditions, and information about estimated by reference to regression equations dietary composition with respect to the relative established for the relationships between prey size proportions of the different prey types consumed. and the dimensions of the particular digestion- Studies of the dietary composition of wild fish resistant hard part in question (e.g. Table 5.1). are almost invariably based on the analysis of However, care must be exercised when interpret- stomach contents, unless one is interested in over- ing the data obtained when using the method all trophic structure, in which case stable isotopes of reconstructed sizes or weights, because it is provide a complementary method. Polunin and very easy to draw incorrect conclusions about die- Pinnegar (Chapter 14, this volume) provide an ex- tary composition and the relative contributions tensive review of the use of stable isotopes. Here I of different prey organisms to the diet of the review three approaches that are usually adopted predator (for discussion see Jobling and Breiby when carrying out analyses of stomach contents: 1986; Jobling 1987; Juanes et al., Chapter 12, this numerical analysis, volumetric analysis and gravi- volume). metric analysis (Windell and Bowen 1978; Bowen It is not usually practicable to quantify rates of 1996). Each approach provides different types of in- food consumption of wild fish by direct observa- formation relating to prey selection and dietary tion, and two different indirect approaches have composition, but none gives any quantitative as- been adopted for the estimation of food intake by sessment of the amount of food consumed. Of the natural fish populations (Windell 1978; Adams and numerical analyses the frequency of occurrence Breck 1990; Bromley 1994). The first approach is gives an estimate of the proportion of the popula- based on calculations of the energy budget and 116 Chapter 5 bioenergetics modelling; the second relates to populations involves examination of the rate at the combination of stomach contents data ob- which food is digested and evacuated from the tained in field surveys with information about stomach. Several models are available for the esti- gastric evacuation rates obtained from laboratory mation of food consumption of wild fish from gas- experiments. tric evacuation data, each model incorporating a Bioenergetics models can be used to estimate set of assumptions about meal frequencies and the food consumption from growth, or vice versa, by form of the curve that describes the evacuation of incorporating information about fish body size, food from the stomach (Windell 1978; Adams and water temperature, diet composition and the Breck 1990; Bromley 1994; Elliott 1994). Few of energy densities of predators and prey. Model out- the models have, however, been rigorously tested puts are, however, subject to errors related to un- under controlled conditions to check their predic- certainties about the accuracy of input variables, tive accuracy (Elliott and Persson 1978; dos Santos insufficient information about the metabolic costs and Jobling 1995). In one approach it is assumed of activity of wild fish, and possible problems with that stomach contents decline exponentially with the extrapolation of physiological rate functions, time and that feeding is continuous between sam- developed in the laboratory, to the field situation. pling times. At the opposite extreme the assump- Models that vary in complexity have been devel- tion is made that a constant amount of food is oped for the estimation of food consumption or evacuated from the stomach per unit time (i.e. growth of natural fish populations, but all have stomach contents decline linearly with time) and their basis in the energy budget equation (Windell that digestion times are long relative to the timing 1978; Weatherley and Gill 1987; Adams and Breck of meals. Both approaches may have merit, since 1990; Hewett and Johnson 1992; Elliott 1994; Brett they may apply to fish that have different feeding 1995; Forseth et al. 2001). habits. For example, fish which feed on small prey A bioenergetics model based on inputs derived organisms or on food that is nutritionally poor from a series of laboratory experiments may pro- (planktivores, herbivores, detritivores) may feed vide good predictions of growth and/or food con- almost continuously, and the pattern of evacua- sumption of fish exposed to conditions towards tion of these species is often found to be exponen- the centre of their environmental tolerance range tial. On the other hand, other species, such as (Whitledge et al. 1998; Wright et al. 1999). How- many piscivores, consume large meals at infre- ever, such models may give erroneous estimations quent intervals, and the evacuation of food from at environmental extremes or when fish are the stomachs of these fish can often be modelled undergoing repeated cycles of food deprivation using a simple linear relationship (Jobling 1986; and refeeding. For example, Wright et al. (1999) Bromley 1994). reported that the models could overestimate the Irrespective of the model applied, gastric evacu- overwintering costs of largemouth bass (Mi- ation rate data must be accurate if food consump- cropterus salmoides) by 20% or more. Further, in tion estimates are to be reliable. Both the rate of tests made on fish undergoing cycles of depriva- evacuation and the form of the curve (e.g. exponen- tion and refeeding, model predictions were in error tial vs. linear) will be influenced by the type of prey, by 25–35% in some instances (Whitledge et al. and, as with other physiological processes, gastric 1998). Given that fish in the wild are unlikely to evacuation rates will be affected by water tem- feed and grow at constant rates and that substan- perature (Jobling 1986; Adams and Breck 1990; tial variations are likely in both the short and long Bromley 1994; dos Santos and Jobling 1995). A term, these test results may have important impli- model that has been used extensively in field cations for field applications of bioenergetics studies of food consumption by fish is the one models. developed by Elliott and Persson (1978): The alternative to the bioenergetics model for --Et Et estimation of food consumption of natural fish CFFtt=-()0ee Et()1- (5.17) Rates of Development and Growth 117

where Ct is the amount of food consumed during lem in fisheries research. Further, even though t hours, F0 and Ft are the stomach content weights much effort has been expended in devising and val- at the beginning and end of a sampling period of idating techniques for age determination of fish t hours, and E is the instantaneous rate of gastric (reviewed by Campana 2001), age determination of evacuation (i.e. gastric evacuation is assumed to be tropical and subtropical species remains problem- exponential). Daily food consumption (C24) is then atic. The second approach to problems related to calculated as the sum of the Ct values obtained fish feeding and growth involves the manipulation for each of the periods within the 24 hours. This of some variable that is suspected of having an model seems to be applicable when feeding is more influence on the growth response. This type of or less continuous within sampling periods, the experiment will usually provide clearer data than amounts of food in the stomach at the start and observational studies, but it may prove extremely end of a period are not the same, and when gastric difficult to carry out adequately controlled ma- evacuation is adequately described by an expo- nipulative experiments in the field. As a conse- nential function. quence, manipulative experiments are usually carried out on captive populations of fish held under artificial conditions, i.e. in hatcheries, large 5.10 CONCLUSIONS culture tanks or net pens, or in a laboratory setting. This, in itself, may create problems related to the Fish, in common with other animals, convert part extrapolation of measurements to field condi- of the organic material they ingest into living bio- tions. Thus, although there exists a large body of mass, but how effectively they are able to accom- information about the various components of the plish this is influenced both by the quantity energy balance equation, collected for many fish and quality of the food they are able to obtain and species subjected to different environmental by the environment to which they are exposed. conditions (for overviews see Brett 1979, 1995; Changes in environmental conditions may affect Weatherley and Gill 1987; Adams and Breck 1990; the growth of fish both via influences on the Jobling 1994, 1997), there remain many unan- absolute and relative amounts of different food swered questions as to how these data can be best types available and via influences on behaviour used to provide information about feeding, growth and physiology. It is within this framework that and the dynamics of production of fish popula- fisheries biologists must attempt to match tions in the wild. their estimates of fish population dynamics and productivity to the food base. Solutions to the problems are sought via a combination of observa- REFERENCES tional studies and manipulative experiments. In the former, data are collected over time to monitor Adams, P.B. (1980) Life history patterns in marine fishes population-related processes, although consider- and their consequences for fisheries management. Fishery Bulletin 78, 1–12. able care must be taken at the planning stage Adams, S.M. and Breck, J.E. (1990) Bioenergetics. In: C.B. to minimize problems associated with sampling Schreck and P.B. Moyle (eds) Methods for Fish Biology. error and bias (Bagenal 1978; Gunderson 1993; Bethesda, MD: American Fisheries Society, pp. Murphy and Willis 1996). Most field studies of fish 389–415. feeding and growth are observational; although Anderson, R.O. and Neumann, R.M. (1996) Length, data from such studies can give insights into popu- weight, and associated structural indices. In: B.R. lation structure, age and growth, and diet, they do Murphy and D.W. Willis (eds) Fisheries Techniques. Bethesda, MD: American Fisheries Society, pp. have shortcomings. For example, analysis of stom- 447–82. ach contents may be used to obtain qualitative in- AOAC (1990) Official Methods of Analysis. Arlington, formation about what fish eat, but the quantitative VA: Association of Official Analytical Chemists. assessment of consumption remains a major prob- Bagenal, T.B. (ed.) (1978) Methods for Assessment of 118 Chapter 5

Fish Production in Fresh Waters. Oxford: Blackwell Beverton, R.J.H. and Holt, S.J. (1957, reprinted 1993) On Scientific. the Dynamics of Exploited Fish Populations. London: Bagenal, T.B. and Tesch, F.W. (1978) Age and growth. In: Chapman & Hall. T.B. Bagenal (ed.) Methods for Assessment of Fish Pro- Blaxter, J.H.S. (1988) Pattern and variety in development. duction in Fresh Waters. Oxford: Blackwell Scientific, In: W.S. Hoar and D.J. Randall (eds) Fish Physiology, pp. 101–36. Vol. XIA. London: Academic Press, pp. 1–58. Balon, E.K. (1975) Terminology of intervals in fish devel- Blaxter, J.H.S. (1992) The effect of temperature on larval opment. Journal of the Fisheries Research Board of fishes. Netherlands Journal of Zoology 42, 336–57. Canada 32, 1663–70. Boeuf, G. and Le Bail, P.-Y. (1999) Does light have an Balon, E.K. (1984) Reflections on some decisive events in influence on fish growth? Aquaculture 177, 129–52. the early life of fishes. Transactions of the American Bowen, S.H. (1996) Quantitative description of diet. In: Fisheries Society 113, 178–85. B.R. Murphy and D.W. Willis (eds) Fisheries Tech- Balon, E.K. (1990) Epigenesis of an epigeneticist: the de- niques. Bethesda, MD: American Fisheries Society, pp. velopment of some alternative concepts on the early 513–32. ontogeny and evolution of fishes. Guelph Ichthyology Brander, K. (1979) The relationship between vertebral Reviews 1, 1–48. number and water temperature in cod. Journal du Balon, E.K. (1999) Alternative ways to become a juvenile Conseil. Conseil permanent international pour or a definitive phenotype (and on some persisting lin- l’exploration de la Mer 38, 286–92. guistic offenses). Environmental Biology of Fishes 56, Brett, J.R. (1979) Environmental factors and growth. 17–38. In: W.S. Hoar, D.J. Randall and J.R. Brett (eds) Fish Baroiller, J.F. and D’Cotta, H. (2001) Environment and sex Physiology, Vol. VIII. London: Academic Press, pp. determination in farmed fish. Comparative Biochem- 599–675. istry and Physiology C 130, 399–409. Brett, J.R. (1995) Energetics. In: C. Groot, L. Margolis and Baroiller, J.-F., Chourrout, D., Fostier, A. and Jalabert, D. W.C. Clarke (eds) Physiological Ecology of Pacific (1995) Temperature and sex chromosomes govern sex Salmon. Vancouver: UBC Press, pp. 2–68. ratios of the mouthbrooding cichlid fish Oreochromis Broekhuizen, N., Gurney, W.S.C., Jones, A. and Bryant, niloticus. Journal of Experimental Zoology 273, A.D. (1994) Modelling compensatory growth. Func- 216–23. tional Ecology 8, 770–82. Baroiller, J.-F., Nakayama, I., Foresti, F. and Chourrout, Bromley, P.J. (1994) The role of gastric evacuation experi- D. (1996) Sex determination studies in two species of ments in quantifying the feeding rates of predatory teleost fish, Oreochromis niloticus and Leporinus fish. Reviews in Fish Biology and Fisheries 4, 36–66. elongatus. Zoological Studies 35, 279–85. Brooks, S. and Johnston, I.A. (1994) Temperature and Baroiller, J.-F., Guigen, Y. and Fostier, A. (1999) Endocrine somitogenesis in embryos of the plaice (Pleuronectes and environmental aspects of sex differentiation in platessa). Journal of Fish Biology 45, 699–702. fish. Cellular and Molecular Life Sciences 55, 910–31. Brown, C.A. and Gruber, S.H. (1988) Age assessment of Bartlett, J.R., Randerson, P.F., Williams, R. and Ellis, the lemon shark, Negaprion brevirostris, using D.M. (1984) The use of analysis of covariance in the tetracycline validated vertebral centra. Copeia 1988, back-calculation of growth in fish. Journal of Fish 747–53. Biology 24, 201–13. Brown, J.J., Ehtisham, A. and Conover, D.O. (1998) Varia- Bayley, P.B. (1977) A method for finding the limits of tion in larval growth rate among striped bass stocks application of the von Bertalanffy growth model and from different latitudes. Transactions of the American statistical estimates of the parameters. Journal of the Fisheries Society 127, 598–610. Fisheries Research Board of Canada 34, 1079–84. Bulow, F.J. (1987) RNA–DNA ratios as indicators of Benfey, T.J. (1992) Hepatic ornithine decarboxylase growth in fish: a review. In: R.C. Summerfelt and G.E. activity during short-term starvation and refeeding in Hall (eds) The Age and Growth of Fish. Ames: Iowa brook trout, Salvelinus fontinalis. Aquaculture 102, State University Press, pp. 45–64. 105–13. Busacker, G.P., Adelman, I.R. and Goolish, E.M. (1990) Benfey, T.J., Saunders, R.L., Knox, D.E. and Harmon, P.R. Growth. In: C.B. Schreck and P.B. Moyle (eds) Methods (1994) Muscle ornithine decarboxylase activity as an for Fish Biology. Bethesda, MD: American Fisheries indication of recent growth in pre-smolt Atlantic Society, pp. 363–87. salmon, Salmo salar. Aquaculture 121, 125–35. Campana, S.E. (2001) Accuracy, precision and quality Beverton, R.J.H. (1992) Patterns of reproductive strategy control in age determination, including a review of the parameters in some marine teleost fishes. Journal of use and abuse of age validation methods. Journal of Fish Biology 41 (Suppl. B), 137–60. Fish Biology 59, 197–242. Rates of Development and Growth 119

Casselman, J.M. (1990) Growth and relative size of calci- Cossins, A.R. and Bowler, K. (1987) Temperature Biology fied structures of fish. Transactions of the American of Animals. London: Chapman & Hall. Fisheries Society 119, 673–88. Couture, P., Dutil, J.-D. and Guderley, H. (1998) Cech, J.J. (1990) Respirometry. In: C.B. Schreck and P.B. Biochemical correlates of growth and condition Moyle (eds) Methods for Fish Biology. Bethesda, MD: in juvenile Atlantic cod (Gadus morhua) from American Fisheries Society, pp. 335–62. Newfoundland. Canadian Journal of Fisheries and Cerrato, R.M. (1990) Interpretable statistical tests for Aquatic Sciences 55, 1591–8. growth comparisons using parameters in the von Craig, J.K., Foote, C.J. and Wood, C.C. (1996) Evidence for Bertalanffy equation. Canadian Journal of Fisheries temperature-dependent sex determination in sockeye and Aquatic Sciences 47, 1416–26. salmon (Oncorhynchus nerka). Canadian Journal of Cerrato, R.M. (1991) Analysis of non linearity effects Fisheries and Aquatic Sciences 53, 141–7. in expected-value parameterizations in the von Crews, D. (1996) Temperature-dependent sex determina- Bertalanffy equation. Canadian Journal of Fisheries tion: the interplay of steroid hormones and tempera- and Aquatic Sciences 48, 2109–17. ture. Zoological Science 13, 1–13. Charnov, E.L. and Berrigan (1991) Evolution of life D’Cotta, H., Fostier, A., Guigen, Y., Govoroun, M. and history parameters in animals with indeterminate Baroiller, J.F. (2001a) Aromatase plays a key role during growth, particularly fish. Evolutionary Ecology 5, normal and temperature-induced sex differentiation of 63–8. tilapia Oreochromis niloticus. Molecular Reproduc- Cho, C.Y., Slinger, S.J. and Bayley, H.S. (1982) Bioenerget- tion and Development 59, 265–76. ics of salmonid fishes: energy intake, expenditure D’Cotta, H., Fostier, A., Guigen, Y., Govoroun, M. and and productivity. Comparative Biochemistry and Baroiller, J.F. (2001b) Search for genes involved in Physiology B 73, 25–41. the temperature-induced gonadal sex differentiation Clarke, A. (1991) What is cold adaptation and how should in the tilapia Oreochromis niloticus. Journal of we measure it? American Zoologist 31, 81–92. Experimental Zoology 290, 574–85. Clarke, A. (1993) Seasonal acclimatization and latitudi- Devries, D.R. and Frie, R.V. (1996) Determination of nal compensation in metabolism: do they exist? age and growth. In: B.R. Murphy and D.W. Willis (eds) Functional Ecology 7, 139–49. Fisheries Techniques. Bethesda, MD: American Clarke, A. (1998) Temperature and energetics: an intro- Fisheries Society, pp. 483–512. duction to cold ocean physiology. In: H.O. Pörtner and dos Santos, J. and Jobling, M. (1995) Test of a food con- R.C. Playle (eds) Cold Ocean Physiology. Cambridge: sumption model for the Atlantic cod. ICES Journal of Cambridge University Press, pp. 3–30. Marine Science 52, 209–19. Conover, D.O. (1990) The relation between capacity for Dutil, J.-D., Lambert, Y., Guderley, H., Blier, P.U., growth and length of growing season: evidence for and Pelletier, D. and Desroches, M. (1998) Nucleic acids implications of countergradient variation. Transac- and enzymes in Atlantic cod (Gadus morhua) tions of the American Fisheries Society 119, 416–30. differing in condition and growth rate trajectories. Conover, D.O. (1992) Seasonality and the scheduling Canadian Journal of Fisheries and Aquatic Sciences of life history at different latitudes. Journal of Fish 55, 788–95. Biology 41 (Suppl. B), 161–78. Elliott, J.M. (1994) Quantitative Ecology and the Brown Conover, D.O. and Present, T.M.C. (1990) Countergradi- Trout. Oxford: Oxford University Press. ent variation in growth rate: compensation for length Elliott, J.M. and Persson, L. (1978) The estimation of of the growing season among Atlantic silversides from daily rates of food consumption for fish. Journal of different latitudes. Oecologia 83, 316–24. Animal Ecology 47, 977–91. Conover, D.O. and Schultz, E.T. (1995) Phenotypic simi- Ferraris, R.P. and Diamond, J.M. (1989) Specific regula- larity and the evolutionary significance of countergra- tion of intestinal nutrient transporters by their dietary dient variation. Trends in Ecology and Evolution 10, substrates. Annual Review of Physiology 51, 125–41. 248–52. Forseth, T., Hurley, M.A., Jensen, A.J. and Elliott, J.M. Conover, D.O., van Voorhees, D.A. and Ehtisham, A. (2001) Functional models for growth and food con- (1992) Sex ratio selection and the evolution of environ- sumption of Atlantic salmon parr, Salmo salar, from a mental sex determination in laboratory populations of Norwegian river. Freshwater Biology 46, 173–86. Menidia menidia. Evolution 46, 1722–30. Gallucci, V.F. and Quinn, T.J. (1979) Reparameterizing, Conover, D.O., Brown, J.J. and Ehtisham, A. (1997) Coun- fitting, and testing a simple growth model. tergradient variation in growth of young striped bass Transactions of the American Fisheries Society 108, (Morone saxatilis) from different latitudes. Canadian 14–25. Journal of Fisheries and Aquatic Sciences 54, 2401–9. Gulland, J.A. (1969) Manual of methods for fish stock 120 Chapter 5

assessment: fish population analysis. FAO Manuals growth of juvenile turbot from different geographic in Fisheries Science 4, 1–154. strains. Journal of Fish Biology 59, 449–54. Gunderson, D.R. (1993) Surveys of Fisheries Resources. Janzen, F.J. and Paukstis, G.L. (1991) Environmental sex New York: John Wiley & Sons. determination in reptiles: ecology, evolution and Gunderson, D.R. (1997) Trade-off between reproductive experimental design. Quarterly Review of Biology 66, effort and adult survival in oviparous and viviparous 149–79. fishes. Canadian Journal of Fisheries and Aquatic Jobling, M. (1986) Mythical models of gastric empty- Sciences 54, 990–8. ing and implications for food consumption. Guy, C.S., Blankenship, H.L. and Nielsen, L.A. (1996) Environmental Biology of Fishes 16, 35–50. Tagging and marking. In: B.R. Murphy and D.W. Willis Jobling, M. (1987) Marine mammal faeces samples as (eds) Fisheries Techniques. Bethesda, MD: American indicators of prey importance: a source of error in Fisheries Society, pp. 353–83. bioenergetics studies. Sarsia 72, 255–60. He, J.I. and Stewart, D.J. (2001) Age and size at first repro- Jobling, M. (1994) Fish Bioenergetics. London: Chapman duction of fishes: predictive models based only on & Hall. growth trajectories. Ecology 82, 784–91. Jobling, M. (1995) Environmental Biology of Fishes. Hewett, S.W. and Johnson, B.L. (1992) Fish bioenergetics London: Chapman & Hall. model 2: an upgrade of a generalized bioenergetics Jobling, M. (1997) Temperature and growth: modulation model of fish growth for microcomputers. Technical of growth rate via temperature change. In: C.M. Wood Report WIS-SG-92-250, Sea Grant Institute, Univer- and D.G. McDonald (eds) Global Warming: Implica- sity of Wisconsin, Madison. tions for Freshwater and Marine Fish. Cambridge: Hirst, B.H. (1993) Dietary regulation of intestinal nutri- Cambridge University Press, pp. 225–53. ent carriers. Proceedings of the Nutrition Society 52, Jobling, M. and Breiby, A. (1986) The use and abuse of fish 315–24. otoliths in studies of feeding habits of marine pisci- Hislop, J.R.G., Harris, M.P. and Smith, J.G.M. (1991) vores. Sarsia 71, 265–74. Variation in the calorific value and total energy con- Jobling, M. and Johansen, S.J.S. (1999) The lipostat, tent of the lesser sandeel (Ammodytes marinus) and hyperphagia and catch-up growth. Aquaculture other fish preyed on by seabirds. Journal of Zoology, Research 30, 473–8. London 224, 501–17. Jobling, M., Meløy, O.H., dos Santos, J. and Christiansen, Hochachka, P.W. (1988) Channels and pumps: determi- B. (1994) The compensatory growth response of the nants of metabolic cold adaptation strategies. Atlantic cod: effects of nutritional history. Aquacul- Comparative Biochemistry and Physiology B 90, ture International 2, 75–90. 515–19. Johnston, I.A. (1993) Temperature influences muscle dif- Hogendoorn, H., Jansen, J.A.J., Koops, W.J., Machiels, ferentiation and relative timing of organogenesis in M.A.M., van Ewijk, P.H. and van Hees, J.P. (1983) herring (Clupea harengus) larvae. Marine Biology 116, Growth and production of the African catfish, Clarias 363–79. lazera (C. & V.) II. Effects of body weight, temperature Johnston, I.A., Vierra, V.L.A. and Hill, J. (1996) Tempera- and feeding level in intensive tank culture. Aquacul- ture and ontogeny in ectotherms: muscle phenotype in ture 34, 265–85. fish. In: I.A. Johnston and A.F. Bennett (eds) Animals Holden, M.J. (1974) Problems in the rational exploitation and Temperature: Phenotypic and Evolutionary of elasmobranch populations and some suggested Adaptation. Cambridge: Cambridge University Press, solutions. In: F.R. Harden Jones (ed.) Sea Fisheries pp. 153–81. Research. London: Elek Science, pp. 117–37. Johnston, I.A., Cole, N.J., Vierra, V.L.A. and Davidson, I. Holden, M.J. and Meadows, P.S. (1962) The structure of (1997) Temperature and developmental plasticity of the spine of the spur dogfish (Squalus acanthias L.) and muscle phenotype in herring larvae. Journal of its use for age determination. Journal of the Marine Experimental Biology 200, 849–68. Biological Association of the United Kingdom 42, Jonassen, T.M., Imsland, A.K., Fitzgerald, R. et al. (2000) 179–97. Geographic variation in growth and food conversion Houlihan, D.F., Mathers, E.M. and Foster, A. (1993) efficiency of juvenile Atlantic halibut related to lati- Biochemical correlates of growth rate in fish. In: tude. Journal of Fish Biology 56, 279–94. J.C. Rankin and F.B. Jensen (eds) Fish Ecophysiology. Jørgensen, E.H., Johansen, S.J.S. and Jobling, M. (1997) London: Chapman & Hall, pp. 45–71. Seasonal patterns of growth, lipid deposition and lipid Imsland, A.K., Foss, A. and Stefansson, S.O. (2001) Varia- depletion in anadromous Arctic charr. Journal of Fish tion in food intake, food conversion efficiency and Biology 51, 312–26. Rates of Development and Growth 121

Kamler, E. (1992) Early Life History of Fish: An mal RNA, actin and myosin heavy chain messenger Energetics Approach. London: Chapman & Hall. RNAs in starved and fed larval Atlantic cod (Gadus Kiess, W., Reich, A., Meyer, K. et al. (1999) A role for morhua). Marine Biology 135, 123–32. leptin in sexual maturation and puberty? Hormone Misra, R.K. (1986) Fitting and comparing several growth Research 51 (Suppl. 3), 55–63. curves of the generalized von Bertalanffy type. Kitano, T., Takamune, K., Kobayashi, T., Nagahama, Y. Canadian Journal of Fisheries and Aquatic Sciences and Abe, S.I. (1999) Suppression of P450 aromatase 43, 1656–59. gene expression in sex-reversed males produced by Murphy, B.R. and Willis, D.W. (eds) (1996) Fisheries rearing genetically female larvae at a high water tem- Techniques. Bethesda, MD: American Fisheries perature during a period of sex differentiation in the Society. Japanese flounder (Paralichthys olivaceus). Journal of Mystkowski, P. and Schwartz, M.W. (2000) Gonadal Molecular Endocrinology 23, 167–76. steroids and energy homeostasis in the leptin era. Laird, L.M. and Stott, B. (1978) Marking and tagging. In: Nutrition 16, 937–46. T.B. Bagenal (ed.) Methods for Assessment of Fish Nammack, M.F., Musick, J.A. and Colvocoresses, Production in Fresh Waters. Oxford: Blackwell Scien- J.A. (1985) Life history of spiny dogfish off the tific, pp. 84–100. Northeastern United States. Transactions of the Larsson, S. and Berglund, I. (1998) Growth and food con- American Fisheries Society 114, 367–76. sumption of 0+ Arctic charr fed pelleted or natural food Nathanailides, C., Lopez-Albors, O. and Stickland, N.C. at six different temperatures. Journal of Fish Biology (1995) Influence of prehatch temperature on the devel- 52, 230–42. opment of muscle cellularity in posthatch Atlantic Lindsey, C.C. (1988) Factors controlling meristic salmon (Salmo salar). Canadian Journal of Fisheries variation. In: W.S. Hoar and D.J. Randall (eds) Fish and Aquatic Sciences 52, 675–80. Physiology, Vol. XIB. London: Academic Press, pp. Ney, J.J. (1990) Trophic economics in fisheries: assess- 197–274. ments of demand–supply relationships between preda- Liu, J., Cui, Y. and Liu, J. (1998) Food consumption and tors and prey. CRC Reviews in Aquatic Sciences 2, growth of two piscivorous fishes, the mandarin fish 55–81. and the Chinese snakehead. Journal of Fish Biology 53, Nomura, T., Arai, K., Hayashi, T. and Suzuki, R. (1998) 1071–83. Effect of temperature on sex ratios of normal and gyno- Love, R.M. (1988) The Food Fishes: Their Intrinsic genetic diploid loach. Fisheries Science 64, 753–8. Variation and Practical Implications. London: Far- Osborne, D.R. and Voogt, P. (1978) The Analysis of rand Press. Nutrients in Foods. London: Academic Press. Løken, S. and Pedersen, T. (1996) Effect of parent type and Patino, R. (1997) Manipulations of the reproductive temperature on vertebrae number in juvenile cod, system of fishes by means of exogenous chemicals. The Gadus morhua (L.), in northern Norway. Sarsia 80, Progressive Fish-Culturist 59, 118–28. 293–8. Patino, R., Davis, K.B., Schoore, J.E. et al. (1996) Sex Løken, S., Pedersen, T. and Berg, E. (1994) Vertebrae num- differentiation of channel catfish gonads: normal bers as an indicator for the recruitment mechanism of development and effects of temperature. Journal of coastal cod of northern Norway. ICES Marine Science Experimental Zoology 276, 209–18. Symposium 198, 510–19. Prein, M., Hulata, G. and Pauly, D. (eds) (1993) Multivari- MacDonald, P.D.M. (1987) Analysis of length–frequency ate methods in aquaculture research: case studies of distributions. In: R.C. Summerfelt and G.E. Hall (eds) tilapias in experimental and commercial systems. The Age and Growth of Fish. Ames: Iowa State ICLARM Study Reviews 20, 1–221. University Press, pp. 371–84. Purdom, C.E. (1993) Genetics and Fish Breeding. Magni, P., Motta, M. and Martini, L. (2000) Leptin: a pos- London: Chapman & Hall. sible link between food intake, energy expenditure, Ratkowsky, D.A. (1986) Statistical properties of alterna- and reproductive function. Regulatory Peptides 92, tive parameterizations of the von Bertalanffy growth 51–6. curve. Canadian Journal of Fisheries and Aquatic McCormick, J.H. and Wegner, J.A. (1981) Responses of Sciences 43, 741–7. largemouth bass from different latitudes to elevated Ricker, W.E. (1979) Growth rates and models. In: W.S. water temperatures. Transactions of the American Hoar, D.J. Randall and J.R. Brett (eds) Fish Physiology, Fisheries Society 110, 417–29. Vol. VIII. London: Academic Press, pp. 677–743. McNamara, P.T., Caldarone, E.M. and Buckley, L.J. Rijnsdorp, A.D., van Leeuwen, P.I. and Visser, T.A.M. (1999) RNA/DNA ratio and expression of 18S riboso- (1990) On the validity and precision of back- 122 Chapter 5

calculation of growth from otoliths of the plaice, Van Pelt, T.I., Piatt, J.F., Lance, B.K. and Roby, D.D. (1997) Pleuronectes platessa L. Fisheries Research 9, 97–117. Proximate composition and energy density of Robinson, J.J. (1996) Nutrition and reproduction. Animal some North Pacific forage fishes. Comparative Reproduction Science 42, 25–34. Biochemistry and Physiology A 118, 1393–8. Rombough, P.J. (1997) The effects of temperature on Weatherley, A.H. and Gill, H.S. (1987) The Biology of Fish embryonic and larval development. In: C.M. Wood Growth. London: Academic Press. and D.G. McDonald (eds) Global Warming: Implica- Weisberg, S. (1993) Using hard-part increment data to tions for Freshwater and Marine Fish. Cambridge: estimate age and environmental effects. Canadian Cambridge University Press, pp. 177–223. Journal of Fisheries and Aquatic Sciences 50, Römer, U. and Beisenherz, W. (1996) Environmental 1229–37. determination of sex in Apistogramma (Cichlidae) Weisberg, S. and Frie, R.V. (1987) Linear models for the and two other freshwater fishes (Teleostei). Journal of growth of fish. In: R.C. Summerfelt and G.E. Hall Fish Biology 48, 714–25. (eds) The Age and Growth of Fish. Ames: Iowa State Schnute, J.T. (1981) A versatile growth model with University Press, pp. 127–43. statistically stable parameters. Canadian Journal of Weltzien, F.-A., Planas, M. and Fyhn, H.J. (1999) Tem- Fisheries and Aquatic Sciences 38, 1128–40. perature dependency of early growth of turbot Schultz, E.T., Reynolds, K.E. and Conover, D.O. (1996) (Scophthalmus maximus L.) and its implications for Countergradient variation in growth among newly developmental progress. Journal of Experimental hatched Fundulus heteroclitus: geographic differences Marine Biology and Ecology 242, 201–10. revealed by common-environment experiments. Whitledge, G.W., Hayward, R.S., Noltie, D.B. and Wang, Functional Ecology 10, 366–74. N. (1998) Testing bioenergetics models under feeding Shearer, K.D. (1994) Factors affecting the proximate regimes that elicit compensatory growth. Transac- composition of cultured fishes with emphasis on tions of the American Fisheries Society 127, 740–6. salmonids. Aquaculture 119, 63–88. Windell, J.T. (1978) Estimating food consumption rates Shine, R. (1999) Why is sex determined by nest tempera- of fish populations. In: T.B. Bagenal (ed.) Methods for ture in many reptiles? Trends in Ecology and Evolution Assessment of Fish Production in Fresh Waters. 14, 186–9. Oxford: Blackwell Scientific, pp. 227–54. Sogard, S.M. and Olla, B.L. (2000) Endurance of simulated Windell, J.T. and Bowen, S.H. (1978) Methods for study of winter conditions by age-0 walleye pollock: effects of fish diets based on analysis of stomach contents. In: body size, water temperature and energy stores. Jour- T.B. Bagenal (ed.) Methods for Assessment of Fish nal of Fish Biology 56, 1–21. Production in Fresh Waters. Oxford: Blackwell Strüssmann, C.A., Moriyama, S., Hanke, E.F., Calsina Scientific, pp. 219–26. Cota, J.C.C. and Takashima, F. (1996) Evidence of ther- Woiwode, J.G. and Adelman, I.R. (1991) Effects of tem- molabile sex determination in pejerrey. Journal of Fish perature, photoperiod, and ration size on growth of Biology 48, 643–51. hybrid striped bass ¥ white bass. Transactions of the Strüssmann, C.A., Saito, T, Usui, M., Yamada, H. and American Fisheries Society 120, 217–29. Takashima, F. (1997) Thermal thresholds and critical Wright, R.A., Garvey, J.E., Fullerton, A.H. and Stein, R.A. period of thermolabile sex determination in two (1999) Predicting how winter affects energetics of age- atherinid fishes, Odontesthes bonariensis and Patago- 0 largemouth bass: how do current models fare? nina hatcheri. Journal of Experimental Zoology 278, Transactions of the American Fisheries Society 128, 167–77. 603–12. Sutton, S.G., Bult, T.P. and Haedrich, R.L. (2000) Rela- Xiao, Y. (1994) Von Bertalanffy growth models with

tionships among fat weight, body weight, water variability in, and correlation between, K and L•. weight, and condition factors in wild Atlantic salmon Canadian Journal of Fisheries and Aquatic Sciences parr. Transactions of the American Fisheries Society 51, 1585–90. 129, 527–38. Xiao-Jun, X. and Ruyung, S. (1992) The bioenergetics of Sæther, B.-S. and Jobling, M. (1999) The effects of ration the southern catfish (Silurus meridionalis Chen): level on feed intake and growth, and compensatory growth rate as a function of ration level, body weight, growth after restricted feeding, in turbot and temperature. Journal of Fish Biology 40, 719–30. Scophthalmus maximus L. Aquaculture Research 30, Youngs, W.D. and Robson, D.S. (1978) Estimation of 647–53. population number and mortality rates. In: T.B. Sæther, B.-S. and Jobling, M. (2001) Fat content in turbot Bagenal (ed.) Methods for Assessment of Fish feed: influence on feed intake, growth and body compo- Production in Fresh Waters. Oxford: Blackwell sition. Aquaculture Research 32, 451–8. Scientific, pp. 137–64. 6 Recruitment: Understanding Density-dependence in Fish Populations

RANSOM A. MYERS

6.1 INTRODUCTION to spawn. Here, I discuss population regulation and variability from the egg stage to first The process of recruitment is the conversion of reproduction. eggs, through a series of density-dependent and There are two principal sources of information density-independent processes, to the fish that on recruitment. First, there are time-series of data, reproduce in the next generation. Many simple containing spawner abundance and subsequent questions concerning recruitment are quantita- recruits. These come primarily from commer- tive questions that require estimates of quantities cially exploited fish stocks (see Fig. 6.1). Fisheries that are very difficult to determine. For example, biologists, with their access to long-term research the question ‘When is the magnitude of re- surveys and extensive catch data, have concen- cruitment determined?’ requires a quantitative trated on evidence from larger geographical scales estimate of the density-dependent and density- that may contain data from more than one popula- independent components of mortality. This tion. Many crucial questions can be addressed question has usually been addressed by correlating with such information but it can seldom be used to estimates of abundance at early ages with later study mechanisms. Second, there are experimen- ages, as reviewed by Cushing (1996). Although tal and observational data on a much smaller scale. such studies are useful, they are very seldom The best of this kind are from coral reef systems, carried out with estimates of measurement error where local populations are studied. Coral reef variability that are crucial for the estimation of the ecologists typically work on relatively common truly important parameters. At the very least, a species, where zygote limitation (i.e. limitation by measure of the estimation error variance of all the number of eggs) is usually viewed as an unim- estimates should be used. portant source of variability in recruitment (Hixon No standard definition of recruitment is and Webster 2002). accepted by all fish ecologists. While all agree that I proceed as follows: I first discuss the maxi- recruitment is the number of fish at some certain mum reproductive rate, the carrying capacity and age or stage, the choice of the age and stage varies. the life stages in which density-dependent and Coral reef fish ecologists often define recruitment stochastic density-independent mortality occurs. as the settlement of pelagic larvae or juveniles This ends with a discussion of future directions for from the plankton (Hixon and Webster 2002; Jones research. I then discuss generalizations that appear et al., Chapter 16, this volume), marine fisheries to be valid across many studies or populations, and biologists usually refer to recruitment as the first on processes that have been shown to be of wide age where fishing occurs, while salmon biologists applicability from various published synthetic usually refer to recruitment as the return of adults analyses. I only briefly review the large literature 124 Chapter 6 on the influence of short- or long-term variation as the compensatory reserve, is the most funda- in the environment on recruitment, as it is well mental of all population parameters. Population covered in a number of recent books, e.g. Cushing biologists and national and international manage- (1996). ment agencies have come to a consensus that the Throughout this chapter, I use the empirical maximum reproductive rate is the appropriate recruitment data that I have compiled and measure of the potential for a population to ‘com- which is freely available from my website (http: pensate’ for fishing, and is used to construct //fish.dal.ca/welcome.html). management targets and limits (Sissenwine and Shepherd 1987; Clark 1991; Mace and Sissenwine 1993; Mace 1994; Myers and Mertz 1998). In this 6.2 THE LINK BETWEEN section, I examine examples of how this parameter SPAWNER ABUNDANCE AND is estimated. When spawners and recruits are SUBSEQUENT RECRUITMENT expressed as absolute estimates of abundance, the maximum reproductive rate can be estimated The spawner–recruit relationship is the basis for directly from the fitted spawner–recruit relation- estimation of the key parameters needed to under- ship. This section explains the most common stand and manage fisheries, e.g. carrying capacity, types of stock-recruitment models and provides maximum reproductive rate and the variability in examples of stock-recruitment data from a variety recruitment. The wide scatter and the relatively of fish populations. ‘flat’ relationship often seen between spawners The relationship between recruitment (R) and and subsequent recruitment misled the naive into spawner abundance (S), perhaps expressed as num- dismissing such data without understanding what bers or biomass, both of which are proxies for egg can be learned from it. Many fisheries biologists production, would take the form believed that the maximum reproductive rate is so high that recruitment would not decline at low RSfS= a (), (6.1) spawner abundance. This belief was due to several where a is the slope at the origin, and f(S) is usually factors. First, it had been observed for many fish a monotonically decreasing function. As such f(S) populations that a reduction in spawner abun- is the relationship between survival and spawner dance was not accompanied by a corresponding abundance. reduction in the production of juvenile fish The parameter a in equation 6.1 defines the (Gilbert 1997). Furthermore, since many individ- scope a population has to compensate for any form ual fish can produce a million eggs or more, this has of increased mortality. For this parameter to be been taken as proof that there can be no shortage of calculated in equation 6.1, it is necessary that the juvenile fish (McIntosh 1899). It was only through units of spawners and recruitment be the same. a very large meta-analysis of over 300 spawner– For species that die after reproduction (e.g. Pacific recruitment datasets that it has been widely ac- salmon, Oncorhynchus spp.) this is simple: one cepted that the production of juvenile fish does in can count the number of female recruits for each fact decline when spawner abundance is greatly female spawner. For species that do not die after reduced (Myers and Barrowman 1996; Myers reproduction, this is more complex. The process 1997). is explained below. The maximum reproductive rate for the popula- The estimation of a requires an extrapolation tion, or equivalently the maximum slope of the because we seldom have abundance estimates at spawner–recruit curve, which is realized as the extremely low population sizes. For this estima- origin is approached where stock size is zero, deter- tion we require a functional form of the density- mines the limits to sustainable fishing and other dependent mortality, f(S), above: sources of anthropogenic mortality that a popula- tion can sustain. The quantity, sometimes known RickerER()= a S e-bS (6.2) Recruitment 125

aS tory density-dependent mortality. There was no Beverton–Holt ER = (6.3) () evidence for depensation in most spawner– 1+ ()SK recruitment time-series because compensatory aS Shepherd ER()= (6.4) mortality appeared to occur throughout the 1+ ()SKg observed range of spawner abundances (Myers et al. 1995a). The clearest evidence for depensation where E(R) is the expected recruitment and -b is occurred when population size was reduced to very the density-dependent mortality. The parameter K low levels, e.g. below 100 spawners. Some species has the same dimensions as the spawners, S, and groups, such as the Clupeidae, contained some may be interpreted as the ‘threshold biomass’ for datasets that were consistent with depensation at the model. For values of biomass S greater than the a moderate abundance of around 20% of unfished threshold K, density-dependent effects dominate. spawner abundance, but it is unclear if this was The Ricker model shows overcompensation, i.e. at due to long-term environmental change or to high spawner abundances recruitment declines. depensation (Myers et al. 1995a; Liermann and The parameter g in the Shepherd model may be Hilborn 1997). We thus conclude that the assump- called the ‘degree of compensation’ of the model, tion that survival, f(S), is a monotonically decreas- since it controls the degree to which the (density- ing function for the observed range of spawner independent) numerator is compensated for by abundances appears to be an adequate approxima- the (density-dependent) denominator (Shepherd tion for most commercial species except at very 1982). low spawner abundances. However, if populations The Ricker and Beverton–Holt formulations are reduced to very low abundance, recovery may are standards in fisheries science, but require some be very slow. For example, the spring-spawning explanation for general ecologists. First, the units herring (Clupea harengus) population in Iceland of R and S may be in biomass as opposed to num- has not recovered since the 1960s, and the cod bers. The reason for this convention is that egg pro- (Gadus morhua) populations in eastern Canada duction in many species is more closely related to have increased at a slower rate than predicted by the biomass of spawners than to the number of the data before the collapse. spawners. The slope, a, determines the highest I now examine several examples of spawner– fishing mortality that can be sustained in a deter- recruit data in which absolute abundances are ministic equilibrium. Note that in these models, estimated. These examples are from a wide variety all density-dependent mortality occurs during the of habitats, include heavily exploited and non- egg, larval or juvenile stage. That is, density- exploited populations, and use a great variety of dependent mortality occurs before the fish ma- techniques in their assessment of the populations. ture. This approximation is consistent with most We begin our examples with a sockeye salmon analysis of data for marine fish (Myers and Cadigan (Oncorhynchus nerka) population because its 1993a). life history is simple to understand: almost all The estimation of the slope at the origin re- females mature at 4 years, spawn once, and die. We quires an extrapolation to zero spawner abun- next consider brook trout (Salvelinus fontinalis), dance, and thus is an approximation. It is clearly a slightly more complex example that was necessary to have at least two individuals for suc- experimentally manipulated to investigate cessful reproduction, and in general almost always density-dependent mortality. The last two exam- more for a population to be self-sustaining. Any ples examine more complex life histories. management system will attempt to keep the population size above the point where such depen- 6.2.1 Sockeye salmon sation occurs. Myers et al. (1995a) and Liermann and Hilborn (1997) examined a large number of For the population of sockeye salmon in the commercially exploited fish stocks for depensa- Adams River, British Columbia, females almost 126 Chapter 6 always live for 4 years. Spawning females are units, in this case millions of females. If we plot the counted in the river; this is the number of one-to-one line, or the replacement line, then it is a spawners on the x-axis (Fig. 6.1). The number of simple matter to see how much the population females produced from this group of spawning would increase if there was no fishing on the females is simply the number that return to spawn population. It can be seen that in most years the 4 years later in the river, plus the number which number of replacement females produced per were harvested on their way to the spawning site. female was approximately five or six. The average Plotting the number of returning females against replacement rate at low abundance, where densi- the number of females that produced them yields a ty-dependent mortality is not important, is given classic spawner–recruit plot. by the slope at the origin of the spawner–recruit Both the spawners and recruits are in the same curve; this quantity is usually termed, by ecolo- gists, the maximum reproductive rate. Based on the fitted spawner–recruitment curves, at low abundance a female can produce approximately 8 8 nine replacements. This allows the population to persist at very high fishing mortalities; typically around 80% of the returning fish from the popula- 6 6 tion are harvested. However, this quantity also defines the limits of exploitation: the population will decline if much more than 80% of the popula- tion is harvested on average, and a lower harvest 4 4 rate would be a much better management strategy. For Pacific salmon, it is relatively easy to calcu-

Resulting spawners late the maximum reproductive rate because the 2 2

Recruits at age 4 (millions) fish are usually relatively easy to count, and be- assuming no fishing (millions) cause complications caused by multiple spawning and complex age structure are avoided. 0 0 0.0 0.5 1.0 1.5 6.2.2 Experimentally manipulated Spawners (millions) brook trout populations Fig. 6.1 Spawner–recruitment data for Adams River sockeye salmon (Oncorhynchus nerka), British Most good population dynamics data come from Columbia, Canada. On the x-axis is the number of heavily exploited populations. We now examine spawning females in the river for any one year. Each seven populations that have experienced very point represents the production of spawners produced little fishing pressure but which have been for each batch of spawners in a given year. These experimentally depleted to determine the nature ‘recruits’ are adults that return to spawn 4 years later. of density-dependent mortality (DeGisi 1994). An The solid and dotted lines are the maximum likelihood experimental approach has clear advantages in fit for the Ricker and Beverton–Holt models respective- ly. The straight dashed line is the replacement line. That that the experimenter can manipulate the popula- is, if a point is above this line, the population should tions at will. increase if no fishing occurs. If 1000 spawning females The seven brook trout populations are from produced 5000 recruits 4 years later, the population small lakes in the Sierra Nevada Mountains in would have increased by five times during the 4-year California. Population sizes were estimated using generation if no fishing had occurred. The solid curved maximum likelihood depletion methods from line is the maximum likelihood estimate of the mean for experimental gill-net surveys. These gave esti- Ricker spawner–recruitment functions under the assumption that the probability distribution for any SSB mates of the number of mature fish present in the is given by a lognormal distribution. The curved dotted population at the time of spawning and an esti- line is the equivalent line for the Beverton–Holt model. mate of the number of fish produced from each Recruitment 127 group of spawners that survived to age 1. In this example, the unit of spawner abundance is the 120 500 number of mature fish present at the time of spawning, and the recruits are the resulting num- 100 400 ber of fish that survive to age 1. In interpreting these data, one problem imme- 80 diately becomes apparent: the units of spawners 300 and recruits are not the same, and thus it is not 60 possible to use these units to estimate the slope at 200 40

the origin as in the sockeye salmon example. To Resulting spawners overcome this problem, the standard procedure is Recruits at age 1 (numbers) 20 100 to convert the units of recruits into the same units assuming no fishing (numbers) as spawners. That is, we estimate the number of 0 0 spawners that would be produced from a given 0 50 100 150 200 number of recruits, if no fishing has occurred. We Spawners (numbers) use the same natural mortality as estimated by the analysis of these data (DeGisi 1994) and ask how Fig. 6.2 Spawner–recruitment data for brook trout many spawners would be produced from the (Salvelinus fontinalis) in the Hell Diver 3 lake in the observed level of recruitment. In this case, the age Sierra Nevada. On the x-axis is the number of spawning adults for any one year. Each point represents the pro- at maturity is 2 and the natural mortality was duction of 1-year-old recruits for each level of observed assumed to be close to 20% a year, although a recruitment on two different scales: on the left-hand side slightly different value was estimated by DeGisi. is the estimated number of trout that survive to age 1, We simplify slightly. Thus 80% of fish that reach and on the right-hand side is the resulting number of age 1 would survive to spawn if 20% died each year. spawners that would be produced if no fishing occurred Once a fish reaches maturity, we need to estimate on the cohort over the lifetime of the cohort. The scale how many times it would spawn on average in on the right-hand side is the scale on the left-hand side multiplied by the average number of times a 1 year old the absence of anthropogenic mortality. For this would be expected to spawn if no fishing occurred (see calculation, we sum the probability that a fish will text for explanation). The straight dashed line is the be alive in each year after it reaches the age of replacement line. The intersection of the Ricker reproduction. This probability is 0.8 for the first spawner–recruitment curve (solid line) with the replace- year after spawning, 0.8 ¥ 0.8 = 0.64 for the second ment line represents the approximate equilibrium for year, and so on. This is known as a geometric the population. The dotted line is the Beverton curve. series, and its sum is 5 in our case. That is, a fish that survived to 1 year of age would spawn an aver- age of four times. This is calculated from the proba- recruits decreases below the equilibrium level, bility of surviving to spawn once, 0.8, multiplied while below this population size the production of by the average number of times the fish would recruits increases. spawn, five. To plot the spawners and recruits in When the population was increased to high the same units, we multiply the number of recruits levels, average recruitment decreased in each of from each age-class by the average number of times the seven experimental lakes. This form of densi- they would spawn over their lifetime. This is the ty-dependent mortality, called overcompensation, scale on the right-hand axis of Fig. 6.2. has rarely been so clearly demonstrated. The maxi- In each of the seven populations, the unex- mum reproductive rate for these populations is ploited population persists around an equilibrium estimated to be around 19 replacement spawners population size. For example, the Hell Diver 3 per spawner from the Ricker model, a value that is population appears to produce an equilibrium similar in all the seven experimental populations when the number of spawners is between 150 and (Myers et al. 1999). 200 (Fig. 6.2). Above this level, the production of It is useful to examine Fig. 6.2 in detail to under- 128 Chapter 6 stand the nature of the extrapolation to the origin. There are three data points where approximately 50 spawners were observed, and we calculate a pro- 600 15000 duction of replacement spawners of around 400 for this case, which results in a slope of around 8. The extrapolation to the origin increases a pattern that 10000 is typical for a curve with overcompensation. If 400 there are observations at low spawner abundances, such as in Fig. 6.1, this is much less of a problem. 200 5000 Also note that the Beverton–Holt model does a Resulting spawners very poor job of fitting these data, and results in an Recruits at age 3 (millions) effectively infinite slope at the origin.

0 0 assuming no fishing (thousand tonnes) 6.2.3 Iceland cod 0 500 1000 1500 Spawners (thousand tonnes) The next example is the cod population from Iceland, based on data collected between 1928 and Fig. 6.3 Spawner–recruitment data for cod from 1995 (Fig. 6.3). The history of this population has Iceland. On the x-axis is the biomass of spawning adults for any one year. Each point represents the production of been reconstructed from detailed catch-at-age recruits for each level of observed spawner biomass on records over this time period and represents one two different scales: on the left-hand side is the recruits of the best-documented fish populations in the in millions of 3-year-old cod, and on the right-hand side world (Schopka 1994). This population has re- is the resulting spawning biomass if no fishing occurred cently shown only a small decline in recruitment on the cohort. To convert the scale on the left-hand axis even though spawning biomass has been reduced to the one on the right-hand axis, multiply the number of to less than one-tenth of what it was 50 years ago. 3-year-old fish by the average spawning lifetime biomass produced by one 3-year-old if no fishing occurred (see More importantly, from the relatively low text for explanation). The straight dashed line is the population level of 200000 tonnes of mature fish, replacement line. The intersection of the spawner– the stock can produce about 200 million 3-year-old recruitment curves with the replacement line represents recruits. Again, we can convert the production of the approximate equilibrium for the population if no these 200 million recruits to the same units as the fishing occurred. spawners. To do this, we ask how much spawning biomass would be produced for each recruit, and then calculate the spawning stock biomass per re- dance can be estimated, and it is also generally the cruit with no fishing, SPRF=0, where the subscript F first age at which it is feasible to harvest. This cal- indicates dependence on fishing mortality, which culation results in the estimate that one 3-year-old is zero for this definition (Mace 1994). This is cod is expected to produce around 24kg of spawn- calculated as ing biomass, if no fishing occurs. That is, the 200 million recruits would be expected to produce • around 4.8 million tonnes of spawning biomass SPR= w l p , (6.5) Faaa=0 Â over their lifetime. The maximum reproductive aa= rec rate can be calculated from the Ricker curve in Fig. where la is the natural survival from the age of 6.3: at low population sizes, a spawner should pro- recruitment to age a, pa is the proportion of duce more than 20 replacements over its lifetime. fish mature at age a, wa is the weight at age a, and This population has sustained a relatively high arec is the age of recruitment, which is 3 years old rate of exploitation, which has typically seen over for Iceland cod. For explicitness, we must state 50% of the fish removed in any one year. Note that that arec represents the first age at which abun- the slope at the origin for the Ricker model is much Recruitment 129 less than that estimated for the Beverton–Holt converted to spawning biomass in the same way model (the dotted line). as for cod described in the last section. These data The cod population in Iceland and the brook show that, at low levels, the replacement level is trout population from lakes in the Sierra Nevada almost 20 spawners per spawner in the absence of Mountains differ in almost every possible way; anthropogenic mortality, similar to that for the however, their maximum reproductive rate is Iceland cod. approximately the same in both cases. The high levels of compensatory reserve esti- mated from the spawner–recruit data are also seen 6.2.4 Striped bass on the east coast from research surveys of the Hudson River popula- tion of striped bass. Pace et al. (1993) showed that of North America the young-of-the-year (YOY) index of striped bass Recently, a detailed assessment of striped bass in the Hudson River was unrelated to the number (Morone saxatilis) on the east coast of North of larvae. This implies either density-dependent America has been completed (National Marine mortality between the early larval stage and the Fisheries Service 1998). The NMFS assessment establishment of the YOY index, or inadequate can be used to estimate spawner recruitment for sampling of the YOY when their abundances are the years 1982–95. The biomass of spawners in- high (see Persson, Chapter 15, this volume). Since creased by a factor greater than five during this the analysis of Pace et al. (1993), spawning stock period, primarily due to reduced fishing mortality. has continued to increase in the Hudson River. The reduction in fishing mortality allowed many Resulting YOY indices have also remained more fish to reach maturity, which also produced a relatively constant despite this increase in subsequent increase in recruitment (Fig. 6.4). stock size (National Marine Fisheries Service The maximum reproductive rate is given by the 1998). slope at the origin of the spawner–recruitment plot. In this case, both the Beverton–Holt and Ricker models give similar estimates of the slope 6.3 GENERALITIES at the origin of the curve. The recruitment was THROUGH META-ANALYSIS

The previous examples of spawner–recruitment 200 relationships come from a wide variety of fish 15000 species with very different life histories and pat- terns of exploitation. For example, the cod popula- 150 tion in Iceland and the brook trout population from 10000 lakes differ in almost every possible way; however, their maximum reproductive rates are very simi- 100 lar. Despite the wide variety in the species exam- ined, one central biological point remains: the 5000 maximum reproductive rate was substantial (i.e. 50 Resulting spawners >1) for all the populations examined. However, the Recruits at age 1 (thousands) maximum reproductive rate is not so high that

assuming no fishing (thousand tonnes) overfishing, particularly of pre-reproductive fish, 0 0 will not cause stock collapse. Since time-series for 0 246810 individual populations are often short and noisy, Spawners (thousand tonnes) we can use meta-analysis to estimate, for example, Fig. 6.4 Spawner–recruitment data for striped bass on the rate of fishing that would lead to stock the east coast of North America. The explanation of the collapse. scales is the same as for Fig. 6.3. Meta-analysis refers to the process of combin- 130 Chapter 6 ing and assessing the findings from several sepa- the maximum reproductive rate divided by the rate research studies that bear upon a common average number of spawnings in an exploited pop- scientific problem (Hedges and Olkin 1985; Coop- ulation, for any of the species examined is usually er and Hedges 1994). The use of statistical methods between one and seven. This number may be less of meta-analysis for research synthesis is now the for some species and more for others, but the rela- standard accepted method for making crucial deci- tive constancy of the annual reproductive rate is an sions in medical treatment, drug evaluations, and unanticipated and very important finding. issues in public health and social policy. Recently, The common belief that there is no relationship it has become a standard approach for evaluating between spawner biomass and recruitment is the critical population parameters needed to un- founded on the notion that the maximum repro- derstand fisheries dynamics (Mace and Sissenwine ductive rate for fish is essentially infinite. This 1993; Myers and Barrowman 1996; Liermann and belief is based on the observations that fecundity Hilborn 1997; Punt and Hilborn 1997; Myers and of fish is often large and that no strong reduction of Mertz 1998; Myers et al. 1999). For its use in exam- recruitment is observed at low spawner abun- ining the effects of fishing on marine ecosystems, dances over the range of the observations. The see Kaiser and Jennings, Chapter 16, Volume 2. problem lies in that little information can be Until recently the formal implementation of gleaned from individual datasets, particularly as meta-analytic methods was hampered by a lack of data at low spawner abundance is sparse. Instead, appropriate statistical methods and of compiled we need to examine a variety of datasets simulta- datasets. These two problems have been largely neously in order to make proper inferences about overcome. The compilations by Mace and cowork- spawner–recruitment relationships (Myers and ers (e.g. Mace and Sissenwine 1993) and by Myers Barrowman 1996; Myers 1997). and coworkers (e.g. Myers et al. 1995b) provide examples of the databases required for a thorough meta-analysis of compensatory reserve. Mace’s 6.4 CARRYING CAPACITY data have been incorporated into the ongoing com- pilation by Myers. Recent advances in statistical A cursory examination of the spawner–recruit software allow for the implementation of complex data for an Icelandic cod stock suggests that the linear and non-linear mixed effects (variance com- carrying capacity is around 200 million 3-year-old ponents) models. These models have been used to recruits (Fig. 6.3). It would be extremely useful to great effect in meta-analysis (Myers et al. 1999). have a theory to explain differences in carrying Meta-analysis of compensatory reserve has capacity, yet none exists. This is perhaps the received a great deal of attention because of its greatest outstanding, and overlooked, problem in usefulness in estimating limits to fishing mor- recruitment. One exception are students of tality, maximum sustainable yield and the shape salmon biology, who have studied this issue with of the spawner–recruit relationship. The broadest some care (e.g. Bradford et al. 1997). Only recently examination of compensatory reserve is an analy- have we begun to study marine species in the same sis of 246 fish populations by Myers et al. (1999). way (Iles and Sinclair 1982; Rijnsdorp et al. 1992; These analyses demonstrate that compensatory Myers et al. 2001). reserve appears to be relatively constant within a No serious ecologist believes that carrying species and within groups of related species. The capacity is constant for any population, as it is analyses employed variance components models certain to change with the environment and the that assume that the log of the standardized slope abundance of predators, parasites and competitors. at the origin of the spawner–recruit curve is a It is crucial that fisheries ecologists begin develop- normal random variable (Searle et al. 1992). This ing a predictive understanding of the variability in analysis suggested a new and unsuspected finding: carrying capacity as it is central to many ecological the maximum annual reproductive rate, which is and management issues. It is clear that a meta- Recruitment 131 analytic perspective will be required, so as to esti- There are very large differences among families mate the variability in carrying capacity both in recruitment variability; however, most of the within and between populations. The first step to data shows recruitment variability with a CV of such an analysis is to standardize carrying capacity around 50% (Fig. 6.5). Surprisingly there exists no on a per unit area basis, so that it can be compared theory to explain this important observation. For across populations. Then it will be possible to example, why is the typical value not around 10% employ non-linear mixed effect models to carry or 100%? We cannot pretend to have a quantitative out a meta-analysis of the carrying capacity for all theory of recruitment until we have explanations regions simultaneously, and to analyse cofactors, of this and similar observations. such as primary productivity and species interac- There are other generalities that emerge about tions, that could cause the carrying capacity to recruitment variability. The spectrum of the time- vary (Myers et al. 2001). I believe that such an series of animal abundance typically contains sig- analysis will greatly improve our understanding of nificant low-frequency variation. Anyone familar the factors that limit carrying capacity, and thus with recruitment time-series will realize that this recruitment. effect is almost always present; however, a few time-series, e.g. recruitment for Icelandic cod, do not show this effect. This may be because this 6.5 VARIABILITY IN species lies within the centre of its range (Myers RECRUITMENT 1991). Several empirical patterns have been pro- posed to help explain the patterns in recruitment Thus far, I have concentrated on the mean relation- variability; these are summarized in Table 6.1. ship between spawner abundance and subsequent recruitment; however, variability in recruitment is often the most notable feature of such data. I 6.6 AT WHAT LIFE-HISTORY now look at recruitment variability using the spawner–recruitment database that I have assem- STAGE DOES DENSITY- bled. I examine two estimates of the variability: DEPENDENT MORTALITY the coefficient of variation (CV) of recruitment and OCCUR? the standard deviations of log residuals from a Ricker spawner–recruitment function (Fig. 6.5). In the spawner–recruit examples, we do not know Maximum likelihood was used to estimate the at what stage density-dependent mortality and parameters of the spawner–recruitment function stochastic density-independent mortality occurs. under the assumption that the variability in This question was first addressed in terms of the recruitment for any spawner abundance was critical period hypothesis that states ‘the numeri- lognormally distributed (see Myers et al. 1995b for cal value of a year class is apparently stated at a details). Similar estimates of the variability were very early age, and continues in approximately the obtained if alternative spawner–recruitment same relation to that of other year classes through- functions were used. Unfortunately, it is very diffi- out the life of the individuals’ (Hjort 1914). This is cult to obtain unbiased estimates of recruitment certainly true to some extent but with the follow- variability. For example, ageing errors will reduce ing limitations: (i) some stochastic mortality after estimated variability in recruitment for catch- and (ii) density-dependent mortality changes the at-age analyses (Bradford 1991), while research relative abundance, but good year-classes remain surveys may overestimate the variability because good. For some populations we can examine how estimation error may be an important source interannual variability changes with age, and use of variability. Thus, these results should serve models to examine when density-dependent and only as a rough guide (see my website for more density-independent mortality occurs. Two con- information). trasting situations can be seen in the data for 132 Chapter 6

0 50 100 150 200 250 0.0 0.5 1.0 1.5 2.0 (a) (b) Acipenseridae A Ammodytidae A Anoplopomatidae A Argentinidae A Bothidae Fig. 6.5 Boxplots by family of (a) the Carangidae Centrarchidae coefficient of variation in recruitment Channichthyidae ignoring the effect of spawner abun- Clupeidae dance and (b) the maximum likeli-

Cyprinidae hood estimate of the standard Engraulidae deviation of the natural log of the Esocidae residuals from fits of the Ricker model Gadidae s assuming lognormal variability in

Haliotidae recruitment (which is an estimate of Hexagrammidae Lophiidae the standard deviation of the density- Lutjanidae Moronidae M independent mortality). Note that not Mugilidae M Nototheiniidae all recruitment series have spawner Osmeridae Pandalidae estimates, so that (a) has more data Pectinidae Penaeidea than (b). The boxplots show the limits Percichthyidae of the middle half of the data (the Percidae white line inside the box represents Plecoglossidae the median). The amount of data is Pleuronectidae shown as the width of the boxes, which are proportional to the square Salmonidae root of the number of data points. The

Sciaenidae notches are the approximate 95% con- fidence intervals of the median. If the Scombridae Scophthalmidae notches on two boxes do not overlap, Scorpaenidae this indicates a difference in a location Serranidae at a rough 5% significance level. The Soleidae Sparidae upper quartile and lower quartile pro- Squalidae vide the outline of the box. Whiskers Stromateidae Trichiuridae are drawn to the nearest value not be- yond 1.5* (interquartile range) from

All the quartiles; points beyond are drawn individually as outliers. Outliers are plotted as points, except for one that falls outside the range of the plot, which has been replaced by an arrow.

brown trout (Salmo trutta) from Black Brows Beck variability at age in cod is caused by strong density- (Elliott 1994) and cod from the North Sea (Fig. 6.6). dependent mortality during the demersal juvenile Very strong density-dependent mortality occurs stage, and is seen in the 17 populations of demersal in the first 6 months after eggs are deposited for fish studied by Myers and Cadigan (1993a,b). brown trout, while mortality that varies among Several observations from the cod data should be years, and is largely independent of density, occurs noted: (i) year-classes that are large soon after set- until maturity at age 3. For brown trout, recruit- tlement remain relatively large for all subsequent ment variability increases with age (Fig. 6.7). surveys, (ii) the variability decreases with age and The pattern is quite different for cod in the (iii) larger spawner abundances are slightly posi- North Sea where the variability in recruitment tively correlated with higher recruitment. The decreases with age. The decrease in recruitment first two observations are clear from virtually all Recruitment 133

Table 6.1 Patterns in recruitment variability.

Pattern Species Reference

CV(R) ~50% Most species This chapter Var(R) ≠ with length of time-series Most species Myers et al. (1995a) Var(R) ≠ at edge of species range Cod, haddock, herring Myers (1991) Var(R) ≠ for greater fecundity (slight effect) 57 species Mertz and Myers (1996), Rickman et al. (2000) Var(R/S) ≠ at low spawner abundance 10 families Myers (2001) Var(R) ≠ with smaller duration of spawning Cod Mertz and Myers (1994) Var(R) ≠ on offshore banks Cod, haddock, American plaice Myers and Pepin (1994) Var(R) Ø at older ages Marine demersal fish Myers and Cadigan (1993a) Var(R) ≠ at older ages Anadromous salmonids Bradford (1995) Var(R) is greater for anadromous than marine Rothschild and DiNardo (1987) Spatial scale of R ~500km 11 marine species Myers et al. (1997a) Spatial scale of R £50km 7 freshwater species Myers et al. (1997a), Bradford (1999)

Var(R), variance of recruitment; CV(R), coefficient of variation in recruitment; Var(R/S), variance of survival for egg production to subsequent recruitment. population dynamics data of demersal fish with to use otolith analysis to examine variation in sur- a planktonic life stage (I return to this in a later vival within and among cohorts (Rice et al. 1987). If section). The last observation is certainly weak a cohort is sampled multiple times, it is possible to for these data, but is a persistent feature of the use otolith microstructure to measure patterns vast bulk of spawner–recruit data (Myers and of daily growth, and hence greatly increase our Barrowman 1996). Note that both of these very understanding of recruitment processes. However, reliable datasets show similar levels of variation at the amount of sampling required to test statistical age 3, with a CV of about 0.5, which we have previ- models with such data can be overwhelming; ously seen to be commonly observed for fish (Fig. careful modelling should precede such a project. 6.5). While the result for demersal fish appears to hold for many populations, the result for trout may 6.7 ESTIMATING DENSITY- not be true for salmonids in general, although DEPENDENT MORTALITY there is sometimes increased variance in demersal FROM LONG-TERM SURVEYS fish due to density-dependent mortality in the juvenile stage (Myers and Cadigan, 1993b; Studies of recruitment have been hampered by the Fromentin et al. 2001). For example, Bradford unwillingness of many biologists to use methods (1995) found that the variance in survival in- that account for the estimation error inherent creased with age for many Pacific salmon species, in estimating the abundance of fish populations. and about half the variability in natural survival With newer methods, it is possible to make better occurred after migration to the sea. It would be of use of the many long-term research surveys of fish great interest to apply the methods that I devel- abundance. These surveys can be used to estimate oped with Cadigan, which include estimation the extent of density-dependent mortality and, if error, to examine the general question of the performed over multiple life-history stages, can be creation of variability in survival by density- used to infer at which life-history stage compensa- independent mortality and the reduction in the tion is occurring. Such an analysis is best carried variance by density-dependent mortality for out using an extension of key-factor analysis that salmonid data. includes estimation error. The original version of A complementary approach to this problem is key-factor analysis assumed that measurements 134 Chapter 6

(a) (b) EGFC Age = 0.5 Age = 0.5 500 80 400 60 300 40 200 100 20 0 0 IYFC Age = 1 Age = 0.75 100 120 80 100 80 60 60 40 40 20 20 0 0 EGFC Age = 1.5 Age = 1.5 60

) 50 2 50 40 40 30 30 20 20

(fish per 60m 10 10 (numbers per hour) R 0 R 0

IYFC Age = 1.75 Age = 2 60 50 30 40 20 30 20 10 10 0 0

EGFC Age = 3 Age = 2.5 8 15

6 10 4 5 2

0 0 0 2000 4000 6000 8000 0 50 100 150 200 250 300 Eggs per 60 m2 Spawners (thousand tonnes)

Fig. 6.6 Recruitment versus spawner abundance at five ages for (a) brown trout and (b) cod. Note that recruitment is estimated from two different surveys for North Sea cod: EGFS is the English Groundfish Survey and IYFS is the International Young Groundfish Survey, which covers a longer time period. Recruitment 135

(a) NNaa+1 =--+exp()td aaaa log N e . (6.6) 1.5 Brown trout Cod Note that the above formulation assumes that density-dependent mortality occurs before the 1.0 stochastic component of density-independent mortality; this should be kept in mind when interpreting the results. 0.5 Myers and Cadigan showed that if estimates of CV of recruitment the observation error variance are available from the sampling variability, or if multiple surveys 0.0 of the same cohort occurred, it was possible to estimate the density-dependent mortality and the (b) variance of the density-independent mortality. 1.5 Furthermore, it is possible to estimate the delayed density-dependent mortality as well. They tested the hypothesis that population variability is 1.0 created and regulated at the juvenile stage for 17 populations of groundfish. Density-dependent mortality was estimated using an extension of 0.5 classical key-factor analysis (Varley and Gradwell 1960) that explicitly included estimation error. Sigma from Ricker model Myers and Cadigan concluded that the juvenile 0.0 stage was crucial for density-dependent popula- 0.5 1.0 1.5 2.0 2.5 3.0 tion regulation in these species, but that the source Age of interannual variability in year-class strength occurs during the larval stage or the very early Fig. 6.7 Changes in (a) the coefficient of variation (CV) juvenile stage. These quantitative estimates con- in recruitment and (b) the standard deviation of the log residuals of the recruitment from a Ricker model at five firm previous analyses that show the strong influ- ages for brown trout and cod. ence of density on mortality in the juvenile demersal stage after settlement to the bottom (Rauck and Zijlstra 1978; Lockwood 1980; Van Der Veer et al. 1990). were made without error; this is not an adequate One major disadvantage of Myers and Cadigan’s approximation for fish population data. method was that it was restricted to the assump- Myers and Cadigan (1993a,b) considered a tion that density-dependent mortality is propor- population in which each generation is surveyed at tional to the logarithm of the initial density. This several stages or ages. Let the number of individu- assumption was needed to obtain linearity of the als entering stage a be Na. Consider the dynamics model equations. However, it should be possible to between two stages, a and a + 1, where the natural consider much more general forms of density- mortality is divided into three components: (i) the dependent mortality using more modern methods, average mortality between the two stages that such as non-linear state space models estimated is independent of density, ta; (ii) the density- using the hierarchical Bayes methods (Gilks et al. dependent component, da, which is assumed to be 1994). proportional to the logarithm of the initial density; Much less is known about density-dependent and (iii) the variable component unrelated to abun- mortality from research surveys of pelagic fish, dance, ea. The dynamics between two stages is probably because of the much greater difficulty of described by: estimating abundance of pelagic fish compared 136 Chapter 6 to demersal species. We know of no analysis for the importance of density-dependent and density- pelagic species similar to the work of Myers and independent mortality at various life stages. Fish- Cadigan that have included the role of estimation eries ecologists have made very inefficient use of error in the statistical models. the excellent survey data available to them; the Long-term surveys of lakes have also generally analysis is difficult and requires the use of subtle shown strong density-dependent mortality during statistical methods. We should make better use of the preadult life stages. For example, long-term such data. surveys using research traps for European perch in Windermere, UK (Mills and Hurley 1990), 6.7.1 Density-dependent mortality trawls for yellow perch in Oneida Lake, New York within and between cohorts (Nielsen 1980) and mark–recapture studies of walleye in Escanaba Lake, Wisconsin (Serns 1982) Spawner–recruit data relies on a regression have demonstrated the importance of density- approach. Another source of evidence, the dependent mortality during preadult life stages. within-cohort data, is usually ignored. Although Density-dependent mortality has been ob- traditional methods of analysis of data within co- served in many long-term research surveys of horts, such as key-factor analysis, are useless for stream fish. This was observed in those cases fisheries work because of the assumption that where cohorts were observed multiple times, for abundance is measured without error, it is possible example brown trout (Elliot 1994), brook trout to extend this type of analysis using modern statis- (McFadden et al. 1967), Atlantic salmon tical methods (Myers and Cadigan 1993a,b). The (Chadwick 1985) and several Pacific salmon basic idea of these methods is not to treat each co- species of the genus Oncorhynchus (Ricker 1954; hort abundance as an unrelated number but to as- Myers et al. 1997b). Strong density-dependent sume that the initial recruitment comes from a mortality and/or growth is expected in species that statistical distribution, usually lognormal, and to have territorial behaviour, which is typical of treat all variation in subsequent ages as also com- stream-dwelling species (Elliott 1994). ing from a statistical distribution. This allows a Extensive surveys of anadromous fish have much deeper understanding of the process. been helpful for establishing the timing of density- Previous work has been limited by rather strict dependent mortality. For example, the monitoring distributional assumptions, for example that re- of the number of anadromous alewife (Alosa cruitment is lognormal. The methods appear to be pseudoharengus) spawners into, and the number reasonably robust to this assumption; perhaps the of seaward migrants out of, Damariscotta Lake, more important assumption is the nature of the Maine showed that a five-fold change in egg pro- functional form of the density-dependent mor- duction resulted in almost no change in recruits; tality. Recent state-space and non-linear time- strong compensation was occurring in the egg or series methods should allow this to be studied in larval stages (Walton 1987). These results are con- the near future. sistent with other studies of Alosa species (Lorda and Crecco 1987), and contrast with those of the demersal fish described above for which strong 6.8 PELAGIC EGG, LARVAL density-dependent mortality existed later in the AND JUVENILE STAGES post-larval stage. In summary, long-term research surveys have The population biology of the pelagic stages provided valuable insights into the nature of popu- remains the most difficult of all problems related lation dynamics and compensation from fishery to understanding recruitment. The primary source independent sources. If estimation error is explic- of variability in recruitment is the enormous itly included in the analysis of the results from interannual variability in density-independent such surveys, they allow quantitative estimates of mortality during the pelagic egg and larval stages Recruitment 137 of marine fish; however, it has been difficult to portance of the match depends on (i) the length of identify the source of this variability (Leggett and time between peaks of the two distributions (t0) DeBlois 1994). Before discussing these issues in and (ii) the width of the distributions, where d and depth, I detail the most influential theory concern- s are measures of the width of the zooplankton and ing interannual variability in larval survival. larval distributions respectively. For example, if the larval distribution has a width of 2 months, a 6.8.1 Match/mismatch hypothesis mismatch of peaks equal to 2 weeks may be in- consequential. Cushing’s idea can be put in a The match/mismatch hypothesis of Cushing simple mathematical model (Mertz and Myers (1969, 1990) has provided considerable insight 1994), which results in powerful testable predic- about year-class success and its variability in tions. Namely, we can calculate how year-to-year marine fish populations. The central idea is that changes in the food consumed by the entire cohort, the closeness of the temporal match between F, are related to those changes in the delay between the abundance peaks of larvae and their planktonic the peak of first-feeding larvae and the peak in their prey controls larval mortality. This may be a result food supply, t0. Since F is a maximum for t0 = 0, we either of the vulnerability of first-feeding larvae to expect that the average value of t0 is zero. The dif- starvation or to the fact that poorly fed larvae grow ference in food consumption, and hence survival slowly and are more susceptible to predation. from the assumption of the hypothesis, can now be Since larval mortality is thought to be very high, simply calculated approximately as the larval stage may be the principal determinant of year-class strength. This picture seems very t2 - 0 (6.7) plausible and it enjoys some empirical support (see ds22+ Cushing 1990 for a review). Figure 6.8, adapted from Cushing (1982), illus- (see Mertz and Myers 1994 for details). That is, a trates the match/mismatch hypothesis. The im- difference in survival is now hypothesized to de-

Match

t0 Eggs Larvae Food

Fig. 6.8 Illustration of a match and a mismatch of larval fish to –40 –20 020 40 60 their planktonic prey (time units are arbitrary). Production of larvae Mismatch P(t) follows egg release and in a t0 match situation (top panel) the Probability density larval production peak closely P(t) Z(t) coincides with the peak of σδ palatable zooplankton (curve designated Z(t)). The match condition is characterized by t0 ~ 0. The representative widths of the –40 –20 020 40 60 larval production and zooplankton peaks are s and d respectively. Time (days) 138 Chapter 6 cline by the square of the difference in timings of reum) predation has been shown to have a signifi- the peaks, t0 divided by the sum of the variances cant impact on year-class strength of yellow perch of the seasonal distribution of first-feeding larvae (Perca flavescens) in Oneida Lake (Mills et al. and zooplankton abundance. This formulation 1987). In contrast, these effects have not been allows a priori quantitative testing of the match/ convincingly demonstrated in marine systems mismatch hypothesis using data that is presently (Leggett and DeBlois 1994). Verity and Smetacek available from regions with regular plankton (1996) have conjectured that predation is a less surveys. dominant influence in marine systems because There are other predictions from the above fish densities appear to be 10–1000 times smaller model that have been tested. Cushing (1990) noted than in freshwater systems (Horn 1972); however, that the effect of variable and unpredictable tim- Horn’s estimate may not be relevant because he ing of the plankton peak will be mitigated if a fish did not estimate densities that should actually stock spreads its spawning effort over a broad occur in the ocean. The orders of magnitude differ- temporal window. This corollary of the match/ ence in the density of larvae clearly suggests that mismatch hypothesis also suggests that there density-dependent mortality may be more impor- might exist a relationship between the width of tant during pelagic larval and juvenile stages in the spawning window, measured by the standard fresh water. deviation of the estimated egg production versus Second, there are fewer sampling difficulties in time curve, for a given stock and its recruitment lakes than in the ocean, although some difficulties variability. Mertz and Myers (1994) were able to are similar, for example larvae avoid the sampling confirm this hypothesis for cod stocks in the device and the avoidance behaviour changes with North Atlantic. age or size. Many marine populations spawn over a wide geographical and temporal window, which 6.8.2 Density-dependent mortality may change among years (Hutchings and Myers in the pelagic stage 1994a), making sampling difficult. For example, the very well studied cod population off Lofoten Density-dependent mortality has been often hy- (Pedersen 1984) actually also spawns outside pothesized as a key factor in population regulation the fjord. The strong advection that occurs for in the pelagic larval stage; however, it has rarely, if marine populations makes estimation very diffi- ever, been convincingly demonstrated for any cult (Helbig and Pepin 1998). The advected larvae marine species (Cushing 1996). For example, at- may die, if they are advected into regions where tempts to demonstrate food limitation, which is they cannot survive (Myers and Drinkwater 1988), the mechanism most often proposed for mortality, or they may survive and return to the spawning have generally yielded ambiguous results at best region at an older age (Polacheck et al. 1992). (Heath 1992; Leggett and DeBlois 1994; Lochmann Error in estimating abundance poses funda- et al. 1997). On the other hand, density-dependent mental limitations to the kinds of questions that mortality in the pelagic stage in freshwater can be addressed. For demersal fish, it is possible species, or pelagic freshwater stages of anadro- to estimate the density-dependent and density- mous species, has been clearly demonstrated in a independent component of mortality between egg number of cases (Walton 1987). and demersal stage, or within the demersal stage. There are several reasons for the difference be- However, reliable estimates of the sources of tween our understanding of freshwater and marine variability within the egg and larval stage are systems (see Jones et al., Chapter 16, this volume). almost impossible except for a very few species First, recruitment variability in fresh water may using enormous sampling effort, as illustrated depend predominantly on biotic influences, par- by the California Cooperative Oceanic Fisheries ticularly the predation by adults of one species on Investigation. juveniles of another. Walleye (Stizostedion vit- Finally, an experimental approach is much Recruitment 139 easier to undertake in freshwater systems than in importance of predicting recruitment. Approaches the ocean. For example, out-planting artificially to this task may involve finding environmental reared larvae or juveniles has been widely used in correlates of recruitment or the field sampling of freshwater systems to study density-dependent pre-recruit life-history stages. Predictions of re- growth and mortality. Whitefish (Corigonidae) has cruitment have been sought through the cor- been experimentally stocked at controlled densi- relates of recruitment with environmental factors. ties in lakes in Finland (Salojarvi 1991), gizzard Wind speed has been proposed as a determinant shad has been experimentally stocked in reser- of recruitment in that storm-driven mixing voirs in Kentucky (Buynak et al. 1992), walleye in can disperse larvae and their prey, reducing ponds in the midwestern USA (Fox and Flowers food availability (Lasker 1975, 1981; Peterman and 1990) and steelhead (Oncorhynchus mykiss) fry Bradford 1987). Larval food supply may also be in- into tributaries of Lake Superior (Close and fluenced by the lag between appearance of larvae Anderson 1992). Each of these studies demon- and the peak of abundance of their prey (Cushing strated strong density-dependence in growth 1990). The intensity of turbulence may control and/or mortality. the frequency of contact between larvae and their Research in this area has concentrated on the prey (Rothschild and Osborn 1988). Larvae may be sources of stochastic density-independent mortal- exported to inhospitable waters by the action of ity caused by lack of food, predation or advection. wind-driven currents or the incursion of Gulf The first two are potentially density dependent, Stream rings (Myers and Drinkwater 1989). For but we have little evidence for marine species that thorough discussions of environmental influ- they are density-dependent in the pelagic stage. ences on recruitment see Fogarty (1993), Wooster The most likely mechanism in the observed cases and Bailey (1989) and Cushing (1996). is food limitation. This has been clearly demon- The complexity of the mechanisms that have strated for pelagic juvenile sockeye salmon, where been proposed can be seen by examining one fac- large year-classes substantially reduce the zoo- tor. Consider the effect of the wind, described by plankton abundance (Hume et al. 1996). Food limi- the vector w, on the survival of pelagic eggs and lar- tation during the freshwater phase of anadromous vae in the Northern Hemisphere. The cross-shelf species that fluctuate greatly in abundance, such wind stress component, tx, will have little effect as the sockeye salmon, will be the most easy to except on the neuston. The long-shelf wind stress demonstrate. component, ty, and the turbulence generated by Although density-dependent mortality is pro- the wind, which is proportional to |w|3 and which bably important for many species in the pelagic will generally be non-linear, may have a plethora stage, it cannot be quantified until a systematic ap- of important effects on survival (Cury and Roy proach is taken to estimation and meta-analysis. 1989; MacKenzie et al. 1994; Dower et al. 1997). These factors may be critical for only short periods 6.8.3 Stochastic density- of the pelagic stage. For example, Ekman transport independent mortality caused by the long-shelf wind stress component in the pelagic stage may decrease survival by transporting larvae off- shore (Bailey 1981) or increase survival by in- The most expensive research efforts in recruit- creasing larval food supply (Bakun 1996). Recent ment have been to understand the creation of vari- theoretical models have shown that small-scale ability in recruitment at pelagic stages of marine turbulence enhances encounter rates between lar- fish. Although progress has been made, the over- val fish and their prey (Rothschild and Osborn all results are meagre; this ambitious research 1988), while earlier work suggested it would de- programme has not resulted in a fully predictive crease survival by dispersing food (Peterman theory. and Bradford 1987). Empirical laboratory and field This research is usually justified because of the observations have helped to clarify the issues 140 Chapter 6

(MacKenzie and Kiorboe 1995); understanding predict recruitment from environmental factors is the physics of the upper ocean allows the relative very limited, even if the principal mechanism de- magnitude of the effects to be estimated, but the termining interannual survival during the egg and final link to resulting recruitment has not been larval stage is well understood, which is seldom established. Similarly, food limitation via Cush- the case. The utility of spending large amounts ing’s match/mismatch hypothesis has been of public research funding to establish predictions investigated and some empirical links from field of recruitment based on environmental indices data demonstrated; again its overall importance is should thus be questioned (Walters and Collie uncertain. It is difficult to produce quantitative 1988). Nevertheless there are broadly two main estimates from any of these models but it is often patterns that emerge from the empirical analysis possible to use data-driven physical oceanographic of such data from a meta-analysis of over 50 stu- models of historic data to quantitatively estimate dies where previously published environment– the effects (Myers and Drinkwater 1988). recruitment correlations had been retested with One approach to understanding the sources of new data (Myers 1998). The two conclusions interannual variability in density-independent were, first, that correlations for populations at the mortality is primarily empirical, by correlations limit of a species’ geographical range have often re- with many environmental correlates. Much of mained statistically significant when re-examined this work has been marred by the extreme unwill- and, second, that a dome-shaped relationship ingness to look at the statistical difficulties of the appears to exist between the recruitment success analyses (reviewed by Myers 1998) and the desire of small pelagic fish in eastern boundary current to blame all fisheries problems on the environ- upwelling systems (Bakun 1996) and the upwelling ment. For example, 5 years after there were claims intensity (optimal environmental window, OEW) that recruitment of ‘northern’ cod could be par- (Cury and Roy 1989). Unfortunately, these envi- tially predicted from the presence of cold water ronment–recruitment correlations do not appear that led to a greater probability of good survival of to be very useful in practice. I recently examined cod larvae, the collapse of the stock was being the 47 environment–recruitment correlations re- blamed primarily on cold water, a claim that did viewed by Shepherd et al. (1984) and found that not stand up to scrutiny (Hutchings and Myers only one was being used in the estimation of 1994b). The claim that poor recruitment of North- recruitment in routine assessments (Myers 1998). ern cod was caused by warm water, and then by cold water, was based on the use of statistical 6.8.4 Demersal stages methods that were at best unreliable. The result- ing comedy of errors, which resulted in the loss Density-dependent mortality probably occurs of around 50000 jobs, should cause any fisheries chiefly in the juvenile demersal stages for coral- scientist to carry out such analyses with more reef fish, shelf demersal species and tidepools care. (Pfister 1996). We know much more about the There are theoretical reasons to doubt that re- mechanisms for density-dependent mortality in cruitment can be predicted with sufficient ac- demersal stages of coral-reef fish than any other curacy from environmental data to be useful for group. The mechanisms have largely been unrav- management (Bradford 1992; Mertz and Myers elled by field experiments that examine how food 1995). The basis of the simulations by Bradford limitation interacts with predation. Predation (1992) and the analytical models of Mertz and mortality can be mediated through territorial Myers (1995) is the distribution of mortality across behaviour or the ‘growth mortality hypothesis’ life-history stages that is typically observed that proposes that greater competition for food in- (Bradford 1992; Bradford and Cabana 1997) and the creases the time required to grow to a size where strength of density-dependent mortality in the ju- they are less vulnerable to predation (Ricker and venile demersal stages (Myers and Cadigan 1993b). Forrester 1948). Forrester (1995a) experimentally When these factors are combined, the ability to demonstrated the second mechanism by artifi- Recruitment 141 cially supplementing food for damselfish dependent mortality during the egg to fry stage in (Dascyllus aruanus). The link between predation brown trout. I find this study particularly inter- and structural complexity was examined by esting because the pattern of density-dependent Beukers and Jones (1998). mortality was not observed in Elliott’s study of Hixon and Carr (1997) translocated 32 live interannual variability. coral-reef heads to a large sand flat so that they The most direct method of assessing com- could manipulate orthogonally the density of prey pensation is by experimental manipulations of and the presence of predators. They demonstrated abundance in a carefully controlled manner. The that density-dependent mortality of juvenile advantage of this approach is obvious: powerful coral-reef fish was dependent upon the presence statistical methods of experimental design can be of two suites of predators: transient piscivores used to construct experiments that can resolve attacking from above and resident piscivores questions that are difficult to answer from obser- attacking from below. If either predator was ab- vational data alone. sent, mortality was entirely density-independent. Similarly, Steele (1997) has shown strong density- dependent mortality using manipulative experi- 6.9 FUTURE RESEARCH ments on temperate juvenile gobies; again, the 6.9.1 The need for density-dependent mortality was primarily due to quantitative theories predation. Forrester (1995b) also showed that juve- nile and adult mortality of gobies was strongly At present, our understanding of recruitment is density-dependent in a series of manipulative ex- limited to conceptual theories and empirical periments. The experimental results that demon- models, meaning spawner–recruitment functions. strate that structural shelter for prey refuges is Although we can estimate biologically meaning- often limiting (Hixon 1991) strongly supports the ful parameters such as carrying capacity and importance of protecting the sea bottom from make quantitative predictions, we lack mechan- destructive fishing practices (Watling and Norse istic models of basic biological and physical 1998). processes. Although we have physical models of The great success of manipulative field experi- some mechanisms (Rothschild and Osborn 1988), ments for coral-reef fish should be emulated they are seldom capable of making predictions by researchers who study other systems. There about recruitment. Thus, few recruitment studies have been some notable successes in other regions. are taken beyond the interpretive stage and can DeGisi’s (1994) whole-lake experiments on brook only claim that observations are consistent with a trout, discussed above, is one of the most ambi- model or a concept. Ultimately, in order for tious. Another experimental approach is to supple- ‘recruitment science’ to make progress, we need ment natural populations with artificially reared mechanistic models that make quantitative pre- competitors. Such large-scale experiments have dictions about the functional form and mean of been undertaken on a marine species, cod, in different processes. Such modelling is difficult, Norway. These have revealed strong density- and unappreciated. dependent mortality in the juvenile stage and My late friend, Gordon Mertz, was a master shown that artificial enhancement of the species of models that yielded testable quantitative pre- in the wild is generally not possible (Nordeide et al. dictions. For example, Mertz developed a model 1994). Density-dependent mortality in this case of Cushing’s match/mismatch hypothesis which appears to be caused by competition for limited predicted that recruitment variability should in- food supply. crease as spawning duration decreases, a predic- There are remarkably few experimental man- tion supported by an analysis of Atlantic cod data ipulations of abundance for stream populations. In (Mertz and Myers 1994). Similarly, he was able to one of the few such studies, LeCren (1965) used en- transform very vague concepts of the relationship closures to demonstrate extremely strong density- between fecundity and recruitment variability 142 Chapter 6 into a quantitative predictive theory (Mertz and petitors (Latto 1992). Therefore, the weight distrib- Myers 1996). I believe that it is this type of model utions of such populations are expected to be posi- that will yield a greater understanding of recruit- tively skewed. Perhaps the reason why such ment, not vague ‘conceptual’, non-parametric or predictions have not been considered to apply well black-box models. to animal populations is the lack of effort to test It is useful for recruitment biologists to exam- these models under field conditions. Sampling ine the role of the fit of a functional model to data. throughout the range of a population under a large For example, Max Planck developed a quantitative variety of densities is beyond the resources of most model for black-body radiation that gave the world field biologists. However, such data are readily the beginnings of quantum theory 100 years ago available for many fish populations from research (Pais 1982). If Max Planck, and the physicists of the surveys. For example, it is a simple matter to plot day, had been satisfied with a non-mechanistic, the length frequencies for Atlantic cod for a year non-parametric fit to the data or if the experimen- of high abundance and low abundance (Fig. 6.9). talist at the University of Berlin had been satisfied For the year-class with high abundance, 1981, the with anything less than the most exacting mea- mean size is reduced and the CV and skewness are surements, we would not now have quantum me- greater compared with a year-class of relatively chanics. Thus, functional models are not the goal low abundance. This relationship holds over all of science; rather they serve as intermediary steps years of the surveys. Atlantic cod shows clear between ever more refined mechanistic models. If territorial behaviour for the first year of life after progress is to be made towards understanding the larval stage (Tupper and Boutilier 1995) and recruitment, we must take models and data more often shows strong density-dependent mortality seriously. (Myers and Cadigan 1993a) and growth. That is, cod may be more like coral-reef fish, at least during 6.9.2 The individual the demersal stage, than most people believe. Such analysis of data from surveys may be very useful in and recruitment understanding recruitment. Modellers of fish populations traditionally as- sumed that all individuals of a given age are identi- 6.9.3 Empirical analysis of cal. However, we know that this is not the case multispecies interactions since competition often takes place between on recruitment individuals; this is the basis of recent work on individual-based models (Cowan et al. 1996; Perhaps the most exciting work in the near Letcher et al. 1996; Huse et al., Chapter 11, Volume future will be the combination of the study of 2). spawner–recruitment relationships with multi- Individual-based models that predict that the species interactions. I predict that the synthesis skewness and the coefficient of variation of weight of these two approaches will yield wonderful new distributions increase with population density discoveries. The massive analysis of many aquatic assume density-dependent growth and unequal communities by Pauly, Christensen and many competition among individuals of a given age coworkers (Christensen and Pauly 1998; Pauly (Kimme 1986). Unequal competition assumes un- and Christensen, Chapter 10, Volume 2) and the equal resource partitioning, which can result from recent extensions to dynamic models (Walters a combination of territorial control of resources, et al. 1997), allied to the approaches outlined here, plastic growth, selective mortality and emigra- could yield remarkable results. tion. Since an animal’s size is highly correlated It is crucial that the interspecific interaction co- to its competitive ability, relevant models predict efficient be comparable over studies. This can be a positive feedback cycle, whereby better com- done by the same standardization as used in the petitors grow larger and become even better com- delayed density-dependent analysis above or by a Recruitment 143

1981 year-class 1985 year-class Age 2 Age 2 25 25 20 skew = –0.15 20 skew = 0.5 CV = 13.9 CV = 8.7 15 mean = 26.8 15 mean = 26.8 10 rec = 439976 10 rec = 217099 5 5 0 0 20304050 20304050

Age 3Age 3

35 skew = 0.6 35 skew = 0.4 30 CV = 10.5 30 CV = 8.7 25 mean = 34.3 25 mean = 37.4 20 rec = 439976 20 rec = 217099 15 15

Frequency 10 10 5 5 0 0 20304050 20304050

Age 4 Age 4

40 skew = 0.4 40 skew = 0.4 CV = 9.1 CV = 7.7 30 mean = 40.1 30 mean = 43.6 rec = 439976 rec = 217099 20 20 10 10 Fig. 6.9 Length frequencies of 0 0 2–4-year-old cod at low (1985) and 20304050 20304050 high (1981) population levels for NAFO Subdivision 2J. Length (cm)

modification of the generalized key factor. All leading approach, even though it is intuitively singles-species spawner–recruitment models appealing. For example, if the reviewer simply are extreme approximations because they ignore examines how often an outcome is statistically species interactions. The complexity of multi- significant, then there will be a strong bias towards species interaction requires a meta-analytic ap- the conclusion that the process or treatment has proach. An excellent informal meta-analysis was no effect (Hedges and Olkin 1985). Furthermore, carried out by Daan (1980) in a comparative study this bias is not reduced as the number of studies in- of species replacement. creases. If a large number of studies or populations are examined, the proportion of studies that yield 6.9.4 Better use of data and the need statistically significant results is approximately for meta-analysis the average power of the test used (Hedges and Olkin 1985). That is, reviewers of research studies When studies are being reviewed, effects are often may assume that they are examining the impor- examined to determine how frequently they are tance of an ecological process, but may only be statistically significant. This can be a very mis- examining the power of the tests used to exa- 144 Chapter 6 mine it. Research synthesis without considering recruitment revolve around multispecies interac- the statistical problems can lead to serious mis- tions; meta-analytic methods to infer these inter- takes. I believe it is a mistake to expect ‘definitive actions from the present data are sorely needed. studies’. Work should focus on obtaining quantitative estimates of the components of density- ACKNOWLEDGEMENTS independent mortality that can be attributed to different causes, such as starvation, predation and I thank J. Gibson, B. MacKenzie and C. Reiss for advection, and quantitative estimates of density- suggestions. dependent mortality at different stages. Results should be presented in such a way that they can be compared among populations and species. The REFERENCES components of the interannual variability in mor- tality can be compared among populations and Bailey, K.M. (1981) Larval transport and recruitment of species (Myers 1995). Similarly, if mortality is pro- Pacific hake, Merluccius productus. Marine Ecology portional to log density, then the coefficient of Progress Series 6, 1–9. density dependence can be compared across popu- Bakun, A. (1996) Patterns in the Ocean: Ocean processes and Marine Population Dynamics. La Jolla, CA: lations (Myers and Cadigan 1993a). Virtually no California Sea Grant College System, National systematic attempts have been made to carry out Oceanic and Atmospheric Administration. these studies. Beukers, J.S. and Jones, G.P. (1998) Habitat complexity modifies the impact of piscivores on a coral reef fish population. Oecologia 114, 50–9. 6.10 CONCLUSIONS Bradford, M.J. (1991) Effects of aging errors on recruit- ment time series estimated from sequential popula- tion analysis. Canadian Journal of Fisheries and Attempts to understand recruitment have domi- Aquatic Sciences 48, 555–8. nated much of fisheries research in the last cen- Bradford, M.J. (1992) Precision of recruitment estimates tury. We have made good progress on some issues. from early life stages of marine fishes. Fisheries Bul- The actual practice of managing fisheries is being letin 90, 439–53. changed into a quantified science, i.e. something Bradford, M.J. (1995) Comparative review of Pacific closer to engineering instead of an art, because we salmon survival rates. Canadian Journal of Fisheries and Aquatic Sciences 52, 1327–38. discovered quantitative generalities and have Bradford, M.J. (1999) Temporal and spatial trends in the developed the tools to apply them, e.g. Bayesian abundance of coho salmon smolts from western North methods. For example, we now have estimates of America. Transactions of the American Fisheries the maximum reproductive rate, which allows the Society 128, 840–6. limits and exploitation to be well understood. Bradford, M.J. and Cabana, G. (1997) Interannual vari- Furthermore, we have a much greater theoretical ability in survival rates and the causes of recruitment and empirical understanding of the nature of variation. In: R.C. Chambers and E.A. Trippel (eds) Early Life History and Recruitment in Fish Popula- recruitment variability (Table 6.1). tions. London: Chapman & Hall, pp. 469–93. Further progress will require improved experi- Bradford, M.J., Taylor, G.C. and Allan, J.A. (1997) Empiri- ments, observations and analytical methods. Al- cal review of coho salmon smolt abundance and the though experiments have elucidated many aspects prediction of smolt production. Transactions of the of post-settlement recruitment processes for coral American Fisheries Society 126, 49–64. reef fish, experimental manipulations need to be Buynak, G.L., Hale, R.S. and Mitchell, B. (1992) Differen- tial growth of young-of-year gizzard shad in several extended to other systems. Recruitment studies Kentucky reservoirs. North American Journal of Fish- have been limited to a few species that are either eries Management 12, 656–62. commercially important or easily studied. The Chadwick, M. (1985) The influence of spawning stock most important outstanding questions concerning on production and yield of Atlantic salmon, Salmo Recruitment 145

salas L., in Canadian rivers. Aquaculture and Fisheries vival and recruitment regulate the abundance of a coral Management 1, 111–19. reef fish. Oecologia 103, 275–82. Christensen, V. and Pauly, D. (1998) Changes in models Fox, M.G. and Flowers, D.D. (1990) Effect of fish den- of aquatic ecosystems approaching carrying capacity. sity on growth, survival, and food consumption by ju- Ecological Applications 8 (Suppl. 1), S104–9. venile walleyes in a rearing pond. Transactions of the Clark, W.G. (1991) Groundfish exploitation rates based American Fisheries Society 119, 112–21. on life history parameters. Canadian Journal of Fish- Fromentin, J.M., Myers, R.A., Bjornstad, O.N., Stenseth, eries and Aquatic Sciences 48, 734–50. N.C., Gjoseter, J. and Christie, H. (2001) Effects of Close, T.L. and Anderson, C.S. (1992) Dispersal, density density-dependent and stochastic processes on the dependent growth, and survival of stocked steelhead regulation of cod populations. Ecology 2, 567–79. fry in Lake Superior tributaries. North American Jour- Gilbert, D.J. (1997) Towards a new recruitment paradigm nal of Fisheries Management 12, 728–35. for fish stocks. Canadian Journal of Fisheries and Cooper, H. and Hedges, L.V. (1994) The Handbook of Re- Aquatic Sciences 54, 969–77. search Synthesis. New York: Russell Sage Foundation. Gilks, W.R., Thomas, A. and Spiegelhalter, D.J. (1994) A Cowan, J.R., Houde, E.D. and Rose, K.A. (1996) Size- language and program for complex Bayesian model- dependent vulnerability of marine fish larvae to ling. The Statistician 43, 169–78. predation: an individual-based numerical experiment. Heath, M.R. (1992) Field investigations of the early life ICES Journal of Marine Science 53, 23–38. stages of marine fish. Advances in Marine Biology 28, Cury, P. and Roy, C. (1989) Optimal environmental win- 1–133. dow and pelagic fish recruitment success in upwelling Hedges, L.V. and Olkin, I. (1985) Statistical Methods for areas. Canadian Journal of Fisheries and Aquatic Sci- Meta-analysis. San Diego, CA: Academic Press. ences 46, 670–80. Helbig, J.A. and Pepin, P. (1998) Partitioning the Cushing, D.H. (1969) The regularity of the spawning sea- influence of physical processes on the estimation of son of some fishes. Journal du Conseil International ichthyoplankton mortality rates. I. Theory. Canadian pour l’Exploration de la Mer 33, 81–97. Journal of Fisheries and Aquatic Sciences 55, Cushing, D.H. (1982) Climate and Fisheries. London: 2189–205. Academic Press. Hixon, M.A. (1991) Predation as a process structur- Cushing, D.H. (1990) Plankton production and year- ing coral reef fish communities. In: P.F. Sale (ed.) The class strength in fish populations: an update of the Ecology of Fishes on Coral Reefs. New York: Academic match/mismatch hypothesis. Advances in Marine Press, pp. 475–508. Biology 26, 249–93. Hixon, M.A. and Carr, M.H. (1997) Synergistic predation, Cushing, D.H. (1996) Towards a Science of Recruit- density dependence, and population regulation in ma- ment in Fish Populations. Oldendorf-Luhe, Germany: rine fish. Science 277, 946–8. Ecology Institute. Hixon, M.A. and Webster, M.S. (2002) Density depen- Daan, N. (1980) A review of replacement of depleted dence in reef fishes: coral-reef populations as model stocks by other species and the mechanisms underly- systems. In: P.F. Sale (ed.) Coral Reef Fishes: Dynamics ing such replacement. Rapports et Procès-verbaux des and Diversity in a Complex Ecosystem. San Diego, Réunions du Conseil International pour l’Exploration CA: Academic Press, pp. 303–25. de la Mer 177, 405–21. Hjort, J. (1914) Fluctuations in the great fisheries of DeGisi, J.S. (1994) Year class strength and catchability Northern Europe. Rapports et Procès-verbaux des Re- of mountain lake brook trout. Master’s thesis, unions du Conseil International pour l’Exploration de University of British Columbia, Vancouver, British la Mer 20, 1–228. Columbia. Horn, M.H. (1972) The amount of space available for Dower, J.F., Miller, T.J. and Leggett, W.C. (1997) The role marine and freshwater fishes. Fisheries Bulletin 70, of microscale turbulence in the feeding ecology of lar- 1295–8. val fish. Advances in Marine Biology 31, 169–220. Hume, J.M.B., Shortreed, K.S. and Morton, K.F. (1996) Elliott, J.M. (1994) Quantitative Ecology and the Brown Juvenile sockeye rearing capacity of three lakes in Trout. Oxford: Oxford University Press. the Fraser River system. Canadian Journal of Fisheries Fogarty, M.J. (1993) Recruitment in randomly varying and Aquatic Sciences 53, 719–33. environments. ICES Journal of Marine Science 50, Hutchings, J.A. and Myers, R.A. (1994a) Timing of cod 247–60. reproduction: interanual variability and the influence Forrester, G.E. (1995a) Factors influencing the juvenile of temperature. Marine Ecology Progress Series 108, demography of a coral reef fish. Ecology 71, 1666–81. 21–31. Forrester, G.E. (1995b) Strong density-dependent sur- Hutchings, J.A. and Myers, R.A. (1994b) What can be 146 Chapter 6

learned from the collapse of a renewable resource? tional Symposium held in Boston, Massachusetts. Atlantic cod, Gadus morhua, of Newfoundland and Bethesda, MD: American Fisheries Society, Vol. 1, pp. Labrador. Canadian Journal of Fisheries and Aquatic 469–82. Sciences 51, 2126–46. Mace, P.M. (1994) Relationships between common bio- Iles, T.D. and Sinclair, M. (1982) Atlantic herring: logical reference points used as threshold and targets of stock discreteness and abundance. Science 215, fisheries management strategies. Canadian Journal of 627–33. Fisheries and Aquatic Sciences 51, 110–22. Kimme, M. (1986) Does competition for food Mace, P.M. and Sissenwine, M.P. (1993) How much imply skewness? Mathematical Biosciences 80, spawning per recruit is enough? In: S.J. Smith, J.J. Hunt 160–71. and D. Rivard (eds) Risk Evaluation and Biological Lasker, R. (1975) Field criteria for the survival of anchovy Reference Points for Fisheries Management. Ottawa: larvae: the relation between inshore chlorophyll maxi- National Research Council, pp. 101–18. mum layers and successful first feeding. Fisheries Bul- MacKenzie, B.R. and Kiorboe, T. (1995) Encounter rates letin 73, 453–62. and swimming behavior of pause-travel and cruise Lasker, R. (1981) Factors contributing to variable recruit- larval fish predators in calm and turbulent laboratory ment of the northern anchovy (Engraulis mordax) environments. Limnology and Oceanography 40, in the California Current: contrasting years, 1975 1278–89. through 1978. Rapports et Procès-verbaux des Réu- MacKenzie, B.R., Miller, T.J., Cyr, S. and Leggett, W.C. nions du Conseil International pour l’Exploration de (1994) Evidence for a dome-shaped relationship be- la Mer 164, 375–88. tween turbulence and larval fish ingestion rates. Lim- Latto, J. (1992) The differentiation of animal body nology and Oceanography 39, 1790–9. weights. Functional Ecology 6, 386–96. McFadden, J.T., Alexander, G.R. and Shetter, D.S. (1967) LeCren, E.D. (1965) Some factors regulating the size of Numerical changes and population regulation in populations of freshwater fish. Mitteilungen Interna- brook trout. Journal of the Fisheries Research Board of tionale Vereinigung für Theoretische und Ange- Canada 24, 1425–59. wandte Limnologie 12, 88–105. McIntosh, W.C. (1899) The Resources of the Sea. London: Leggett, W.G. and DeBlois, E. (1994) Recruitment in ma- C.J. Clay and Sons and Cambridge University Press. rine fishes: is it regulated by starvation and predation Mertz, G. and Myers, R.A. (1994) Match/mismatch in the egg and larval stages? Netherlands Journal of Sea predictions of spawning duration versus recruitment Research 32, 119–34. variability. Fisheries Oceanography 3, 236–44. Letcher, B.H., Rice, J.A., Crowder, L.B. and Rose, K.A. Mertz, G. and Myers, R.A. (1995) Estimating the pre- (1996) Variability in survival of larval fish: disentan- dictability of recruitment. Fisheries Bulletin 93, 657– gling components with a generalized individual-based 65. model. Canadian Journal of Fisheries and Aquatic Sci- Mertz, G. and Myers, R.A. (1996) Influence of fecundity ences 53, 787–801. on recruitment variability of marine fish. Canadian Liermann, M. and Hilborn, R. (1997) Depensation Journal of Fisheries and Aquatic Sciences 53, 1618–25. in fish stocks: a hierarchic Bayesian meta-analysis. Mills, C.A. and Hurley, M.A. (1990) Long-term studies on Canadian Journal of Fisheries and Aquatic Sciences the Windermere populations of perch (Perca fluvi- 54, 1976–85. atilis), pike (Esox lucius) and Arctic charr (Salvelinus Lochmann, S.E., Taggart, C.T., Griffin, D.A., Thompson, alpinus). Freshwater Biology 23, 119–36. K.R. and Maillet, G.L. (1997) Abundance and condition Mills, E.L., Forney, J.L. and Wagner, K.J. (1987) Fish pre- of larval cod (Gadus morhua) at a convergent front on dation and its cascading effect on the Oneida Lake western bank, Scotian Shelf. Canadian Journal of Fish- food chain. In: W.C. Kerfoot and A. Sih (eds) Predation: eries and Aquatic Sciences 54, 1461–79. Direct and Indirect Impacts on Aquatic Communi- Lockwood, S.J. (1980) Density-dependent mortality in ties. London: University Press of New England, pp. O-group plaice (Pleuronectes platessa L.) populations. 113–18. Journal du Conseil International pour l’Exploration Myers, R.A. (1991) Recruitment variability and range of de la Mer 39, 148–53. three fish species. Northwest Atlantic Fisheries Orga- Lorda, E. and Crecco, V.A. (1987) Stock–recruitment rela- nization Scientific Council Studies 16, 21–4. tionship and compensatory mortality of American Myers, R.A. (1995) Recruitment of marine fish: the shad in the Connecticut River. In: M.J. Dadswell, R.J. relative roles of density-dependent and density- Klauda, C.M. Moffitt, R.L. Saunders, R.A. Rulifson and independent mortality in the egg, larval, and juvenile J.E. Cooper (eds) Common Strategies of Anadromous stages. Marine Ecology Progress Series 128, 308–9. and Catadromous Fishes: Proceedings of an Interna- Myers, R.A. (1997) Comment and reanalysis: paradigms Recruitment 147

for recruitment studies. Canadian Journal of Fisheries Myers, R.A., MacKenzie, B.R., Bowen, K.G. and and Aquatic Sciences 54, 978–81. Barrowman, N.J. (2001) What is the carrying capacity Myers, R.A. (1998) When do environment–recruit corre- for fish in the ocean? A meta-analysis of population lations work? Reviews in Fish Biology and Fisheries 8, dynamics of North Atlantic cod. Canadian Journal of 285–305. Fisheries and Aquatic Sciences 58, 1464–76. Myers, R.A. (2001) Stock and recruitment: generaliza- National Marine Fisheries Service (1998) 26th Northeast tions about maximum reproductive rate, density Regional Stock Assessment Workshop. NEFSC Ref. dependence and variability using meta-analytic Document 98/03. Woods Hole, MA: NMFS. approaches. ICES Journal of Marine Science 58, Nielsen, L.A. (1980) Effect of walleye (Stizostedion vit- 937–51. reum vitreum) predation on juvenile mortality and re- Myers, R.A. and Barrowman, N.J. (1996) Is fish recruit- cruitment of yellow perch (Perca flavescens) in Oneida ment related to spawner abundance? Fisheries Bulletin Lake, New York. Canadian Journal of Fisheries and 94, 707–24. Aquatic Sciences 37, 11–19. Myers, R.A. and Cadigan, N.G. (1993a) Density- Nordeide, J.T., Fossa, J.H., Salvanes, A.G.V. and dependent juvenile mortality in marine demersal fish. Smedstad, O.M. (1994) Testing if year-class strength Canadian Journal of Fisheries and Aquatic Sciences of coastal cod, Gadus morhua L., can be determined at 50, 1576–90. the juvenile stage. Aquaculture and Fisheries Manage- Myers, R.A. and Cadigan, N.G. (1993b) Is juvenile natural ment 25, 101–16. mortality in marine demersal fish variable? Canadian Pace, M.L., Baines, S.B., Cyr, H. and Downing, J.A. (1993) Journal of Fisheries and Aquatic Sciences 50, 1591–8. Relationships among early life stages of Morone ameri- Myers, R.A. and Drinkwater, K. (1988) Ekman transport cana and Morone saxatilis from long-term monitoring and larval fish survival in the Northwest Atlantic. Bio- of the Hudson River estuary. Canadian Journal of Fish- logical Oceanography 6, 45–64. eries and Aquatic Sciences 50, 1976–85. Myers, R.A. and Drinkwater, K.F. (1989) The influence of Pais, A. (1982) Subtle is the Lord: the Science and the Gulf Stream warm core rings on recruitment of fish in Life of Albert Einstein. Oxford: Oxford University the Northwest Atlantic. Journal of Marine Research Press. 47, 635–56. Pedersen, T. (1984) Variation of peak spanning of Arcto- Myers, R.A. and Mertz, G. (1998) The limits of exploita- Norwegian cod (Gadus morhua L.) during the time tion: a precautionary approach. Ecological Applica- period 1929–1982 based on indices estimated from tions 8 (Suppl. 1), S165–9. fishery statistics. In: E. Dahl, D.S. Danielssen, D.S. Myers, R.A. and Pepin, P. (1994) Recruitment variability Moksness and E. Solemdal (eds) The Propagation of and oceanographic stability. Fisheries Oceanography Cod Gadus morhua L. Arendal, Norway: Flødevigen 3, 246–55. rapportser, pp. 301–16. Myers, R.A., Bridson, J. and Barrowman, N.J. (1995a) Peterman, R.M. and Bradford, M.J. (1987) Wind speed and Summary of worldwide stock and recruitment data. mortality rate of a marine fish, the northern anchovy Canadian Technical Report of Fisheries and Aquatic (Engraulis mordax). Science 235, 354–6. Sciences 2024, 327. Pfister, C.A. (1996) Consequences of recruitment varia- Myers, R.A., Barrowman, N.J., Hutchings, J.A. and tion in an assemblage of tidepool fishes. Ecology 77, Rosenberg, A.A. (1995b) Population dynamics of ex- 1928–41. ploited fish stocks at low population levels. Science Polacheck, T., Mountain, D., McMillan, D., Smith, W. 269, 1106–8. and Berrien, P. (1992) Recruitment of the 1987 year Myers, R.A., Mertz, G. and Bridson, J.M. (1997a) Spatial class of Georges Bank haddock (Melanogrammus ae- scales of interannual recruitment variations of marine, glefinus): the influence of unusual larval transport. anadromous, and freshwater fish. Canadian Journal of Canadian Journal of Fisheries and Aquatic Sciences Fisheries and Aquatic Sciences 54, 1400–7. 49, 484–96. Myers, R.A., Bradford, M.J., Bridson, J.M. and Mertz, G. Punt, A. and Hilborn, R. (1997) Fisheries stock assess- (1997b) Estimating delayed density-dependent mortal- ment and decision analysis: the Bayesian approach. Re- ity in sockeye salmon (Oncorhynchus nerka): a meta- views in Fish Biology and Fisheries 7, 35–65. analytic approach. Canadian Journal of Fisheries and Rauck, G. and Zijlstra, J.J. (1978) On the nursery aspects Aquatic Sciences 54, 2449–63. of the Wadden Sea for some commercial fish species Myers, R.A., Bowen, K.G. and Barrowman, N.J. (1999) and possible long-term changes. Rapports et Procès- The maximum reproductive rate of fish at low popula- verbaux des Réunions du Conseil International pour tion sizes. Canadian Journal of Fisheries and Aquatic l’Exploration de la Mer 172, 266–75. Sciences 56, 2404–19. Rice, J.A., Crowder, L.B. and Holey, M.E. (1987) Explo- 148 Chapter 6

ration of mechanisms regulating larval survival in Réunions du Conseil International pour l’Exploration Lake Michigan bloater: a recruitment analysis based de la Mer 185, 255–67. on characteristics of individual larvae. Transactions of Sissenwine, M.P. and Shepherd, J.G. (1987) An alterna- the American Fisheries Society 116, 703–18. tive perspective on recruitment overfishing and bio- Ricker, W.E. (1954) Stock and recruitment. Journal of the logical reference points. Canadian Journal of Fisheries Fisheries Research Board of Canada 11, 559–623. and Aquatic Sciences 44, 913–18. Ricker, W.E. and Forrester, R.E. (1948) Computation Steele, M.A. (1997) Population regulation by post- of fish production. Bulletin of the Bingham Oceano- settlement mortality in two temperate reef fishes. graphic Collection, Yale University 11, 173–211. Oecologia 112, 64–74. Rickman, S., Dulvy, N., Jennings, S. and Reynolds, J. Tupper, M. and Boutilier, R.G. (1995) Effects of con- (2000) Recruitment variation related to fecundity in specific density on settlement, growth and post- marine fishes. Canadian Journal of Fisheries and settlement survival of a temperate reef fish. Journal of Aquatic Sciences 57, 116–24. Experimental Marine Biology and Ecology 191, Rijnsdorp, A.D., Beek, F.A.V., Flatman, S. et al. (1992) 209–22. Recruitment of sole stocks, Solea solea (L.), in the Van Der Veer, H.W., Pihl, L. and Bergman, M.J.N. (1990) Northwest Atlantic. Netherlands Journal of Sea Re- Recruitment mechanisms in North Sea plaice (Pleu- search 29, 173–92. ronectes platessa). Marine Ecology Progress Series 64, Rothschild, B.J. and DiNardo, G.T. (1987) Comparison of 1–12. recruitment variability and life history data among Varley, G.C. and Gradwell, G.R. (1960) Key factors in marine and anadromous fishes. Journal du Conseil population studies. Animal Ecology 29, 299–401. International pour l’Exploration de la Mer 1, 531–46. Verity, P.G. and Smetacek, V. (1996) Organism life cycles, Rothschild, B.J. and Osborn, T.R. (1988) Small-scale predation, and the structure of marine pelagic ecosys- turbulence and plankton contact rates. Journal of tems. Marine Ecology Progress Series 130, 277–93. Plankton Research 10, 465–74. Walters, C., Christensen, V. and Pauly, D. (1997) Struc- Salojarvi, K. (1991) Compensation in a whitefish (Core- turing dynamic models of exploited ecosystems from gonus lavaretus) population maintained by stocking in trophic mass-balance assessments. Reviews in Fish Lake Kallioinen, Northern Finland. Finnish Fisheries Biology and Fisheries 7, 39–172. Research 12, 65–76. Walters, C.J. and Collie, J.S. (1988) Is research on environ- Schopka, S.A. (1994) Fluctuations in the cod stock off mental factors useful to fisheries management? Iceland during the twentieth century in relation to Canadian Journal of Fisheries and Aquatic Sciences changes in the fisheries and environment. ICES Ma- 45, 1848–54. rine Science Symposia 198, 175–93. Walton, C.J. (1987) Parent–progeny relationship for es- Searle, S.R., Casella, G. and McCulloch, C.E. (1992) Vari- tablished population of anadromous alewives in a ance Components. New York: John Wiley & Sons. Maine lake. American Fisheries Society Symposium 1, Serns, S.D. (1982) Influence of various factors on den- 451–4. sity and growth of age-0 walleyes in Escanaba Lake, Watling, L. and Norse, E.A. (1998) Effects of mobile Wisconsin, 1958–1980. Transactions of the American fishing gear on marine benthos: disturbance of the Fisheries Society 111, 299–306. seabed by mobile fishing gear. A comparison to forest Shepherd, J.G. (1982) A versatile new stock–recruitment clearcutting. Conservation Biology 12, 1180–97. relationship for fisheries, and the construction of Wooster, W.W. and Bailey, K.M. (1989) Recruitment sustainable yield curves. Journal du Conseil Interna- of marine fishes revisited. In: K.J. Beamish and C.A. tional pour l’Exploration de la Mer 40, 67–76. McFarlane (eds) Effects of Ocean Variability on Shepherd, J.G., Pope, J.G. and Cousens, R.D. (1984) Varia- Recruitment and an Evaluation of Parameters Used in tions in fish stocks and hypotheses concerning their Stock Assessment Models. Ottawa, Canada: Depart- links with climate. Rapports et Procès-verbaux des ment of Fisheries and Oceans, pp. 153–9. 7 Life Histories of Fish

J.A. HUTCHINGS

7.1 INTRODUCTION sion of traits most closely related to fitness, such as 7.1.1 What is a life history? age and size at maturity, number and size of off- spring, and the timing of the expression of those All fish differ in the means by which they propa- traits throughout an individual’s life. Life-history gate genes to future generations. Some mature dur- theory is perhaps most often used to predict the age ing their first year of life at a comparatively small and size at which an organism first reproduces and size, while others delay initial reproduction for the effort it allocates to reproduction at that and tens of years, by which time they can be extremely subsequent ages, often measured as the numbers large. Some produce large numbers of small off- and size of offspring or proportional commitment spring while others produce few numbers of of energy reserves. comparatively large offspring. Some spawn several In a practical sense, life-history theory can be times throughout their lives while others die after used to predict how changes to abiotic and biotic reproducing just once. It is differences such as environments, ultimately through their effects these, usually most pronounced among families on age-specific survival and fecundity, influence and orders, that predicated the seminal research on selection on the fundamental ‘decisions’ that fish life histories by Svärdson (1949) on offspring genotypes face concerning reproduction. For ex- size and number, Alm (1959) on the influence of ample, will a particular environmental perturba- growth rate on age and size at maturity, and tion, such as increased mortality among adults, Beverton and Holt (1959) on associations among favour genotypes that mature early or late in life? growth rate, mortality and longevity. Life-history How might selection influence reproductive effort research thus addresses the tremendous vari- allocated to the behavioural, physiological and en- ability that exists in reproductive strategies among ergetic components of reproduction, each of which species and, perhaps more importantly from an will have possible consequences to future growth ecological and management perspective, among and survival? In addition, will genotypes that par- populations within species and among individuals tition their reproductive effort among few large within the same population. offspring be favoured over those that produce Life histories can be defined as the means by many small offspring? which individuals, or more precisely genotypes, Implicit within any prediction based on life- vary their age-specific expenditures of reproduc- history theory is the assumption that natural tive effort in response to physiologically, environ- selection favours those genotypes whose age- mentally, ecologically and genetically induced specific schedules of survival and fecundity gener- changes to age-specific survival and age-specific ate the highest per-capita rate of increase, or fecundity. As such, life histories reflect the expres- fitness, relative to other genotypes in the same 150 Chapter 7 population. This genotypic rate of increase is recruitment is discussed by Myers (Chapter 6, this usually expressed by r, the intrinsic rate of natural volume). increase. The analytical link between life history A population’s rate of increase, rpop, is simply a and fitness, or rate of increase, is often expressed by function of the mean and variance of the fitness of the discrete-time version of the Euler–Lotka the various genotypes that comprise that popula- equation, in which the fitness of genotype i, i.e. r(i), tion, i.e. m (r(i)) and s2 (r(i)), underscoring the neces- can be calculated from sity of understanding the factors that influence

x=t life-history variation among individuals within 1 = Â limix ()x ()exp()- rix() (7.1) the same population. When estimates of rpop x=a are available over time for a single population, where lx represents the probability of surviving the influence of environmental stochasticity, i.e. from birth until age x, mx represents fecundity at unpredictable environmental perturbations that age x, and age ranges from maturity (a) until death proportionately affect age-specific survival and

(t). The age-specific parameters lx and mx are com- fecundity of all individuals equally, can be incor- monly termed ‘age-specific survival’ and ‘age- porated into an estimate of intrinsic rate, such that specific fecundity’ respectively. the stochastic estimate of r, rs, is

2 7.1.2 Why study life histories? rrs =-d 12()sr (7.2)

2 The study of life histories can be said to be at the where rd is the deterministic estimate of r and sr is heart of most research in ecology and evolution. the variance in r through time (Lande 1993). The This is because a life history is based on the two important point to note here is that increased envi- most fundamental components of life and the two ronmental variability in demographic parameters most important metrics of fitness: survival and re- will negatively influence a population’s long-term production. If a genotype is to be successful evolu- growth rate, such that stochastic estimates of r tionarily, in that it is able to contribute its genes to will always be less than deterministic estimates. succeeding generations at approximately the same Similarly, if fitness estimates of a single geno- rate as other genotypes in the same population, type, or more broadly a specific life-history then it must survive to ages at which it can repro- strategy, are available over time, the fitness associ- duce, and it must contribute to the production of ated with that genotype or life history should be offspring that will themselves survive to repro- estimated as the geometric mean of the fitness duce and produce viable offspring in the future. estimates. Use of the geometric mean makes Notwithstanding its relevance to ecology and implicit the fact that fitness is determined by a evolution, life-history research has an integral role multiplicative process: the total number of de- to play in the management and conservation of scendants left by an individual after n generations fish. This is because of the fundamental link that depends on the product of the number surviving to exists between individual life histories and popu- reproduce in each generation (Seger and Brock- lation rate of increase. Among other things, rate of mann 1987). The geometric mean fitness of geno- increase is inextricably linked to a population’s type i after n generations can be calculated as ability to sustain different levels of exploitation, to ()1 n increase after being reduced to low abundance and ri() ¥ ri() ¥ ri() ¥ ri() ¥¥... ri() [()1234()()() ()n ] to withstand introductions of exotic species with (7.3) which it might interact as competitor, predator or prey. Because the population rate of increase Note that the geometric mean is strongly influ- through recruitment is so important to under- enced by low values. Importantly, then, the more standing exploited populations, it is given special variable a set of values, the lower the geometric attention by fisheries biologists. The topic of mean. Thus, selection should act to reduce the Life Histories 151 variance in genotypic/individual fitness over gen- Gadidae, Pleuronectidae and Molidae (Bond 1996; erations, even if this entails a ‘sacrifice’ of the ex- Helfman et al. 1997; Wootton 1998). pected fitness within any one generation (Roff However, notwithstanding the considerable 1992). Such a ‘bet-hedging’ life-history strategy in life-history variability that exists among species, fish, manifested as the spreading of reproductive there can be extraordinary variability in life histo- risk across space and time, would include multiple ries within and among populations of the same spawning events within a single breeding season, species. The Atlantic salmon, Salmo salar, is breeding with multiple mates and multiple among the most variable of vertebrates and pro- reproductive bouts throughout an individual’s vides an excellent example of the breadth in life- life. Myers (Chapter 6, this volume) deals speci- history variation that can exist within species fically with the spatial and temporal variability of (Hutchings and Jones 1998). Among populations, recruitment. age at maturity differs ten-fold, ranging from 1 year for males in several North American and European 7.1.3 How variable are fish populations (Hutchings and Jones 1998) to as life histories? much as 10 years for anadromous females in northern Québec (Power 1969). Similarly, size at When one thinks of life-history variability in fish, maturity varies at least 14-fold, from less than it is the extraordinary differences among species 7cm for males in Newfoundland (Gibson et al. that initially come to mind. Age at maturity 1996) to more than 1m among females in Nor- ranges from weeks in the tropical cyprinodont way’s River Vosso (Huitfeldt-Kaas 1946). Number Nothobranchius sp. that inhabit temporary pools of eggs per female varies from tens (Gibson et al. (Simpson 1979) to an average of 35 years for spiny 1996) to tens of thousands (Fleming 1996), while dogfish, Squalus acanthias, off British Columbia mean egg diameter ranges from at least 4.5mm in (Saunders and McFarlane 1993). Size at maturity a non-anadromous Newfoundland population ranges from 8–10mm and tenths of a gram in the (Gibson et al. 1996) to 6.9mm in Canada’s Res- smallest goby, Trimmatom nanus (Winterbottom tigouche River (Canadian Department of Fisheries and Emery 1981), to several metres and thousands and Oceans, unpublished data). of kilograms in the largest elasmobranch, the Although the spatial scale of life-history vari- whale shark, Rhincodon typus (Helfman et al. ability among populations of many fish spans a 1997). The elliptical myxinid eggs are typically 2– broad geographical range, one does not require a 3cm in length (Martini et al. 1997; J.A. Hutchings, broad geographical range to detect evolutionarily unpublished data) while oviparous chondri- important variation in life history. This is an im- chthyan eggs are typically 6–7cm in diameter in portant point. One implication for studies of evo- egg cases up to 30cm in length (Bond 1996). Teleost lutionary ecology is that great distances, as much egg sizes range widely. Among the smallest are as several degrees of latitude, are not a prerequisite those produced by the surfperch, Cymatogaster for selection to effect adaptive variation among aggregata (0.3mm diameter; Kamler 1992; populations. For fishery managers and conserva- Wootton 1998), and the lightfish Vinciguerria spp. tion biologists, one implication is the possibility (0.5mm; Helfman et al. 1997); the largest among that sustainable harvest levels and minimum oviparous fishes are those produced by salmonids viable population sizes can differ significantly (2.5–10mm; Tyler and Sumpter 1996) and by among populations across very small geographical mouth-brooding marine catfishes (14–30mm; scales. Tyler and Sumpter 1996; Helfman et al. 1997). Small-scale life-history variation among popu- Number of offspring per female can vary from as lations of brook trout, Salvelinus fontinalis, on few as two in viviparous sharks, such as the sand Cape Race, Newfoundland, serves to illustrate tiger, Odontaspis taurus (Helfman et al. 1997), to this point (Ferguson et al. 1991; Hutchings 1991, several millions in species belonging to the 1993a,b, 1994, 1996, 1997). Selecting the most di- 152 Chapter 7 vergent populations, Freshwater River and Cripple reproductive survival because of the smaller body Cove River are separated by approximately 9km. size typically associated with earlier maturity Brook trout inhabiting these rivers are unex- within a population. By contrast, the primary cost ploited, do not interbreed, do not differ in density of delaying initial spawning is the increased risk of and are not influenced by interspecific competi- death prior to reproduction. The primary fitness tion or predation (Ferguson et al. 1991; Hutchings advantage to delaying maturity in fish is the larger 1993a). Despite these similarities, females in initial body size attained by individuals when they Freshwater River mature, on average, more than a first reproduce. Body size can be postitively associ- full year earlier (3.1 years) than those in Cripple ated with survival (Hutchings 1994), fecundity Cove River (4.2 years) at a five-fold smaller size (Wootton 1998), fertilization success (Hutchings (~11g). Indeed, the minimum lengths at maturity, and Myers 1988; Jones and Hutchings 2001, 2002), 62mm for males and 70mm for females among ability to provide parental care (Wiegmann and Freshwater River trout, may be the smallest Baylis 1995), probability of attracting mates (Foote recorded for this species. Regarding reproductive 1988) and ability to acquire and defend nest sites effort, relative to Cripple Cove River females, the (van den Berghe and Gross 1989). smaller Freshwater River females allocate more If one assumes that natural selection acts on than twice as much body tissue to gonads, produce age-specific expectations of producing future off- significantly more eggs and produce 40% larger spring (Fisher 1930), age at maturity can be predict- eggs, all the comparisons having been corrected ed from the mean and the variance in juvenile and for body size (Hutchings 1991, 1993a, 1996). The adult survival rates, that is, the survival preceding significant population differences in life history and following age at maturity respectively. Reduc- between these populations have been attributed tions in age at maturity are assumed to be favoured to the consequences of environmental differences with decreases in the ratio of adult to juvenile sur- in food supply, and possibly habitat, to age- vival and with increases in the variance in adult specific survival probabilities, fecundity and survival relative to the variance in juvenile sur- growth rate (Hutchings 1991, 1993a,b, 1994, 1996, vival (Cole 1954; Gadgil and Bossert 1970; Schaffer 1997). 1974a,b). Intuitively these predictions make sense. As external mortality (that unassociated with reproduction) at potentially reproductive 7.2 INFLUENCE OF ages increases, selection would be expected to SURVIVAL AND GROWTH favour those individuals (genotypes) that repro- duce prior to those ages, thus increasing their prob- RATE ON AGE, SIZE AND ability of contributing genes to future generations. REPRODUCTIVE EFFORT A similar argument can be made for environmen- AT MATURITY tal perturbations that increase the variance in sur- 7.2.1 Age and size at maturity vival at potentially reproductive ages, increased variance in survival being associated with in- Age at maturity reflects an evolutionary compro- creased uncertainty in an individual’s persistence. mise between the benefits and costs to fitness These predictions are generally presumed to of reproducing comparatively early or late in life hold true for fish. Leggett and Carscadden (1978) (Kozlowski and Uchmanski 1987; Roff 1992; examined population differences in age at ma- Charlesworth 1994). Benefits associated with turity among five populations of American shad, early maturity include increased probability of Alosa sapidissima, from Florida, USA north to surviving to reproduce and an increased rate of New Brunswick, Canada. They found that males gene input into the population, resulting in re- and females in the northern populations, for which duced generation time. However, early maturity they presumed juvenile mortality to be more vari- can result in reduced fecundity and/or post- able than that in southern populations, matured as Life Histories 153 much as 11 and 14% older, respectively, than their biologically meaningful time period, such as gonad southernmost counterparts. Similarly, Hutchings development, movement/migration to spawning and Jones (1998) reported a negative correlation grounds, reduction of feeding prior to and con- between temporal variance in adult survival and comitant with reproduction, mate competition, age at maturity in anadromous Atlantic salmon. nest construction, or parental care. In fish, a Reznick et al. (1990), documenting selection re- common surrogate of reproductive effort is the sponses to predator-induced changes to mortality gonadosomatic index (GSI), i.e. the weight of in guppies, Poecilia reticulata, found age at matu- an individual’s gonads relative to that individual’s rity among males and females to be 17 and 7% total body weight. Among species, GSI ranges from higher, respectively, in the environment charac- as little as 0.2% in male Tilapia spp. (Helfman et al. terized by high juvenile mortality relative to that 1997) and 0.5% in male white sticklebacks, in the environment characterized by compara- Gasterosteus sp. (C.-A. Smith and J.A. Hutchings, tively low juvenile mortality. Fox and Keast (1991), unpublished data) to as much as 47% in female comparing life histories of pumpkinseed sunfish, European eels, Anguilla anguilla (Kamler 1992). Lepomis gibbosus, populations subjected to either Additional metrics of reproductive effort have high or low overwinter mortality, documented 1–2 included direct estimates of energy or lipid losses year reductions in age at maturity among males associated with reproduction (e.g. Dutil 1986; and females in the high-mortality environments. Jonsson et al. 1991; Hutchings et al. 1999), magni- And age at maturity in bluegill sunfish, Lepomis tude and duration of parental care behaviours (e.g. macrochirus, populations exposed to high juvenile Hinch and Collins 1991), size-specific fecundity predation is reported greater than that in popula- (Leggett and Carscadden 1978; Koslow et al. 1995), tions that experience comparatively low juvenile migration distance to spawning grounds (Schaffer mortality (Belk 1995). and Elson 1975) and changes to lean body mass Notwithstanding widespread acceptance of the (Giloolly and Baylis 1999). legitimacy of these tests, none included estimates Changes to reproductive effort in fish are of adult survival relative to juvenile survival nor thought to be associated with adaptive responses attempted to exclude costs of reproduction from to age at maturity. For example, a decline in estimates of adult mortality. Nonetheless, in one survival during potentially reproductive ages study that attempted to fulfil these expectations, (i.e. adult survival) relative to that during pre- there is support for the aforementioned pre- reproductive ages (i.e. juvenile survival) is predict- dictions of life-history theory. Among brook ed to favour genotypes that increase reproductive trout populations on Cape Race, Newfoundland, effort and, in addition, that mature earlier in life Canada, female age at maturity ranges from 3.1 to (Gadgil and Bossert 1970; Law 1979). Evidence 4.2 years among populations whose adult:juvenile of such a direct response to selection has been survival ratios range between 0.16 and 3.19 respec- forthcoming from Reznick et al.’s (1990) work on tively (Hutchings 1993a,b). Evidence of a similar Trinidad guppies: 30–60 generations after a shift in association between adult:juvenile survival ratio predator-induced mortality from adults to juve- and age at maturity has recently been docu- niles, guppies responded to the presumed increase mented for brook trout populations elsewhere in in the ratio of adult to juvenile survival by reducing Newfoundland (Adams 1999). reproductive allotment and by increasing age at maturity. Life-history comparisons among 7.2.2 Reproductive effort populations have also supported these predictions concerning effort and maturity. Comparing Following Hirshfield and Tinkle (1975), reproduc- pumpkinseed sunfish from five populations that tive effort can be defined as the proportion of total experienced either high or low levels of overwinter energy devoted to the physiological and behav- mortality, Fox and Keast (1991) found males and ioural aspects of reproduction, measured across a females inhabiting high-mortality environments 154 Chapter 7 to mature earlier and to have significantly higher 7.2.3 Costs of reproduction GSIs than those inhabiting low-mortality environ- ments. Among brook trout populations in New- A central tenet of life-history theory is that the be- foundland, declines in the ratio of adult to juvenile havioural, physiological and energetic correlates survival are associated with earlier maturity and of reproduction exact some sort of cost to future re- increase in reproductive effort, as approximated by productive success in the form of reduced survival, GSI (Hutchings 1993a). A negative association be- fecundity and/or growth. Firstly, relative to a non- tween age at maturity and reproductive effort, reproductive individual of age x, reproduction at measured as GSI and size-specific fecundity, has age x can reduce the probability of survival to age also been reported for yellow perch, Perca x + n, where n might represent any unit of time flavescens, populations in central Alberta, Canada from minutes to years. Secondly, reproduction can (Jansen 1996). directly reduce an individual’s future ability to Reproductive effort has been predicted to de- produce offspring. High energetic costs expended cline with increased variance in juvenile survival at age x might leave a fish with insufficient energy relative to that of adult survival. Murphy (1968) in- reserves to produce the same number of offspring directly tested this hypothesis by correlating a at age x + n. The third cost is a consequence of proxy for recruitment variability with longevity in the energy allocated to the behavioural and physio- five clupeiforms, finding a positive association be- logical demands of reproduction at the expense of tween the two, suggesting that increased variation energy that would otherwise have been allocated in juvenile survival favours genotypes that reduce to somatic growth. Reduced future size-at-age reproductive effort and, as a consequence, live concomitant with a reduction in growth rate, cou- longer lives. However, additional tests for positive pled with the positive association typically ob- associations between recruitment variability and served between fecundity and body size in fish longevity in fish have not yielded significant corre- (Wootton 1998), results in a reduction in potential lations (e.g. Roff 1991). Leggett and Carscadden fecundity, such that an individual reproducing at (1978) compared North American shad popula- age x will produce fewer eggs at age x + n than an tions and found that northern populations, which individual that did not reproduce at age x. they suggested experienced highly variable The seminal work in this area began with in- juvenile mortality, produced fewer eggs per unit ductive arguments for why reproductive costs body size and had a higher incidence of repeat should exist and theoretical assessments of the spawning than southern populations, presumed life-history consequences thereof (e.g. Fisher 1930; to experience comparatively low variation in Cole 1954; Williams 1966; Gadgil and Bossert juvenile mortality. 1970). The literature turned to a consideration of Reproductive effort has also been suggested to whether costs indeed existed and, if they did, what increase with age and size (Williams 1966; Gadgil were the most appropriate means of measuring and Bossert 1970). Based on his analysis of 31 fish them (e.g. Reznick 1985). The primary challenge stocks, Roff (1991) found support for Charlesworth now lies in how to best estimate costs in natural and León’s (1976) conditions under which repro- populations, and to identify the primary physical ductive effort would be expected to increase with and biological environmental factors that effect age. Evidence of a positive association between variability in the expression of costs between some metric of reproductive effort and age has sexes, among individuals, among years, among also been documented in brook trout (Hutchings generations and among populations. 1993a, 1994) and smallmouth bass, Micropterus To quantify reproductive costs precisely, one dolomieui (Gillooly and Baylis 1999). would manipulate reproductive effort for a spe- cific genotype and then document the genotype- specific survival and/or fecundity consequences of those changes to effort. For most fish, excepting Life Histories 155 parthenogenetic species, these are experimental Lawrence River, Canada, to be more than twice criteria that cannot be met, notwithstanding the that of their breeding counterparts. Using a similar difficulty of assessing the degree to which a cost comparison of survival probabilities, Hutchings quantified under experimental conditions in the (1994) documented evidence of survival costs of re- laboratory is likely to reflect costs experienced production in three Newfoundland populations of in the wild. Nonetheless, if we are to have any brook trout that appeared to increase with age but success in applying life-history theory to our to decline with body size. Indirect evidence of a understanding of the ecology, management and negative association between reproductive cost conservation biology of fish, some attempt must and body size has also been recorded for the sand be made to quantify reproductive costs. Reznick goby, Pomatoschistus minutus (Lindström 1998), (1985) discusses the strengths and weaknesses as- and Atlantic silverside, Menidia menidia (Schultz sociated with various means of estimating costs in and Conover 1999). As a practical tool for quantify- general, while Hutchings (1994) discusses these ing survival costs in the field, this method holds a methods in relation to the study of fish. great deal of promise and is critical if life-history One means of estimating costs, widely used in theory is to be applied to specific management and avian and Drosophila sp. life-history research (e.g. conservation issues. Although one will inevitably Pettifor et al. 1988; Chapman et al. 1993), is to ex- be unable to control all the factors potentially in- perimentally manipulate some metric of present fluencing individual quality, costs estimated by effort and then document the survival and/or such a technique should tend to underestimate fecundity consequences thereof. Examples of such rather than overestimate costs, if one assumes that manipulation in the fish literature are compara- the reproductive individuals are the highest quali- tively rare and are limited to species that exhibit ty individuals (Hutchings 1994). parental care (Balshine-Earn 1995). Notwith- Perhaps the most widely acknowledged cost in standing the attraction of such an experimental fish associated with reproduction at age x is the re- approach, it has its limitations. For example, costs duction in growth rate immediately prior to age x, estimated by manipulation of brood size will un- and thence to ages x + n, that effects a concomitant derestimate actual costs because they exclude the decrease in fecundity, generated by the diversion of physiological and behavioural costs associated energy from somatic growth to the behavioural, with producing the brood. Also, by artificially ma- energetic and physiological demands of reproduc- nipulating the size of brood to which a genotype tion (Bell 1980; Hutchings 1993a; Balshine-Earn was ‘expecting’ to provide care, one risks either 1995; Wootton 1998; Jobling, Chapter 5, this underestimating or overestimating costs if some volume). form of individual optimization (Pettifor et al. It has been argued that the most informative 1988) or adaptive phenotypic plasticity (Hutchings method for measuring a cost of reproduction 1996; Jonsson et al. 1996) exists in the population would be to quantify correlated genetic responses under study. estimated from a selection experiment (Reznick One potentially informative method of esti- 1985). One such experimental protocol would be mating survival costs of reproduction is to quanti- to select only those individuals having either low fy the difference in survival probabilities between or high fecundity to breed at age x, to repeat this ages x and x + n for individuals that reproduced at selection regime over several generations and then age x and those that did not. To reduce the effects of to determine whether there has been a correlated individual differences in aspects of quality, such as response to this selection regime in the form of a body condition, on such an analysis (cf. Reznick negative association between, for example, fecun- 1985), one should control for both size and age. For dity at age x and fecundity at age x + n. Despite their example, Dufresne et al. (1990) found the survival experimental limitations, intentional selection among non-breeding 1- and 2-year-old threespine experiments in the field do have the potential sticklebacks, Gasterosteus aculeatus, in the St to provide valuable insight into how reproductive 156 Chapter 7 costs vary with changes to age-specific survival sitism have not been unequivocally documented, and fecundity. An example of intentional selection if such links exist, increased parasitism could is a change to age-specific mortality effected by be implicated in acute and chronic reproductive natural predators (Reznick et al. 1990); an example costs, given observations that parasites have been of unintentional selection is a change to age- shown to have a wide range of effects on fitness specific mortality effected by fishing. (Barber and Poulin, Chapter 17, this volume). Many reproductive costs in fish can be attrib- These include reduced growth rate (Walkey and uted to the loss of lipids and proteins associated Meakins 1970), inhibited sperm development with various physiological and behavioural corre- (Sinderman 1987), lower gonad weight as in the lates of reproduction, such as gonad production common goby, Pomatoschistus microps (Pam- (Wootton 1998), mate competition (Grantner and poulie et al. 1999), decreased fecundity as found in Taborsky 1998) and parental care (Lindström 1998; Pacific hake, Merluccius productus (Adlerstein Mackereth et al. 1999). These energetic demands and Dorn 1998), and reduced probability of attract- can be considerable, particularly when one ing mates as observed for threespine stickleback compares the energetic losses of reproductive (Milinski and Bakker 1990). The energetic de- individuals relative to those of non-reproductive mands of reproduction might also effect acute, individuals during the same time interval. For short-term survival costs if individuals experience example, post-reproductive Arctic char, Salveli- an increased risk of predation because of reduced nus alpinus, possess 35–46% less energy than non- locomotion, reduced vigilance and/or increased reproductive char in spring (Dutil 1986). Rijnsdorp feeding rate. and Ibelings (1989) reported energy losses among Ecological constraints of reproduction in fish reproductive North Sea plaice, Pleuronectes might result in reduced future survival probabili- platessa, to be three to five times that of non- ties, notably in the short term. Two primary means reproductive plaice. The source of energy losses by which these might be effected are through differs between sexes, with non-gonadal reduc- increased risk of predation and increased risk of tions typically being greater among males than physical injury. The former can result from behav- females (Jonsson et al. 1991; Hutchings et al. 1999), iours associated with attracting mates (Houde and presumably as a consequence of the behaviours Endler 1990) or caring for young (Pressley 1981). associated with mate competition. The magnitude Physical injury, primarily as a result of mate com- of total energy losses attributed to reproduction petition, can also be expected to negatively influ- also differs between sexes, but apparently not in a ence short- and long-term survival probabilities similar manner among, and potentially within, (Hutchings and Myers 1987; Fleming 1996). species. Energetic losses have been reported to be Reproductive costs in fish might also result greater for females in Arctic char (Jørgensen et al. from genetic trade-offs between correlated charac- 1997) and plaice (Rijnsdorp and Ibelings 1989), ters of fitness, although evidence of such antago- equal in Atlantic salmon (Jonsson et al. 1991), yet nistic pleiotropy has not been reported for fish. greater among males in brook trout (Hutchings et al. 1999). 7.2.4 Growth rate Energetic constraints associated with parental care alone have been reported to negatively influ- Growth rate is of major importance to life-history ence future growth and fecundity (e.g. Balshine- evolution, notably in organisms that continue to Earn 1995; Wiegmann and Baylis 1995). Energetic grow after attaining maturity, because of its deter- losses might also effect survival and fecundity mination of body size at age and because of the costs if reproduction is associated with immune positive associations that can exist between body system deficiencies, leading to an increased risk size and various metrics of fitness, such as fecun- of infection or parasitism. Although direct links dity (Wootton 1998), egg size (Kamler 1992), sur- between reproduction and infection risk or para- vival (Hutchings 1994), fertilization success Life Histories 157

(Hutchings and Myers 1988; Thomaz et al. 1997; Reductions in age at maturity with increases in Jones and Hutchings 2001, 2002), parental care individual growth rate have been repeatedly docu- (Sargent et al. 1987; Wiegmann and Baylis 1995) mented in fish (e.g. Alm 1959; Hutchings 1993a; and migration distance (Schaffer and Elson 1975). Fox 1994; Trippel et al. 1995; Godø and Haug 1999). Once the physiological minimum reproductive The capacity of individuals to respond in such size has been attained, an individual’s maturation a manner, and the rapidity with which these strategy is predicted to depend not only on the changes can occur within populations, relative to consequences to present and future survival of generation time, suggests that this life-history maturing at various ages but also on the conse- response is phenotypically plastic, the magnitude quences to present and future fecundity. of which would depend on the shape of the norm of A trade-off between present and future reaction for age at maturity and on the magnitude fecundity, effected by declines in growth rate con- of environmental change in growth rate. Two ex- comitant with maturity, formed the basis of initial amples are gadids reported by Trippel et al. (1997) predictions of how growth rate might influence and yellowtail flounder, Pleuronectes ferruginea, life history (Gadgil and Bossert 1970; Schaffer by Walsh and Morgan (1999). Using empirical 1974a,b; Bell 1980). Schaffer (1974b), for example, age-specific survival, fecundity and growth rate predicted that environments that allowed for data for Newfoundland populations of brook trout, increased growth during potentially reproductive Hutchings (1996) explored the individual conse- ages should favour delayed maturity and increased quences to fitness of maturing at various ages in reproductive effort. Although Schaffer and Elson’s response to changes in growth rate. The resultant (1975) positive association between age at maturi- fitness functions supported the prediction that a ty and growth rate at sea in Atlantic salmon ap- reaction norm describing a negative association peared to support this hypothesis, the correlation between growth rate and age at maturity can did not differ significantly from zero upon reanaly- represent an adaptive response to environmental sis (Myers and Hutchings 1987a). change. However, for one population, fitness was Nonetheless, there is reason to believe that maximized by maturing as early in life as possible, Schaffer’s (1974b) predictions hold merit. Hutch- regardless of growth rate. Selection for such a flat, ings (1993a) suggested distinguishing the growth non-plastic reaction norm can be expected when rate experienced during the juvenile stage from the probability of realizing the fitness benefits of that experienced during the adult stage. He pre- delayed maturity (increased fecundity for females, dicted that increases in juvenile growth rate rela- increased access to mates for males) is relatively tive to adult growth rate should favour increases in low (see Section 7.5). reproductive effort and reductions in age at matu- rity. These predictions were supported by empiri- 7.2.5 Life-history invariants cal data on Newfoundland populations of brook trout (Hutchings 1993a). Fox (1994) tested this The objective of most life-history research in fish hypothesis and reported that pumpkinseed sun- is to test hypotheses of adaptive and non-adaptive fish introduced into a previously fishless pond ex- variation to explain the extraordinary life-history perienced a higher ratio of juvenile to adult growth variability expressed within and among species. rate, matured at an earlier age and had a signifi- An alternative approach has been to determine cantly higher GSI relative to those in the original whether there is a constancy, or invariance, among population. Positive associations between the life-history traits that may reflect adaptive life- ratio of adult to juvenile growth rate and age at history processes of a very broad and general maturity also exist for Newfoundland populations nature. This search for constancy amidst diversity of ouananiche (S. salar; Leggett and Power 1969) was very much evident when quantitative studies and Swedish populations of brown trout (S. trutta; of patterns of growth, maturation and longevity in Näslund et al. 1998). fish first appeared in the late 1950s. Alm (1959), for 158 Chapter 7 example, sought patterns among growth rate, age Another invariant that appears to have some con- and size at maturity from experiments he had con- sistency among taxa is La/L•, which Jensen (1997) ducted on Swedish populations of brown trout. suggests is 0.66 for fish, implying that all fish Using data primarily from commercially impor- mature at a length that is approximately two- tant marine species, Beverton and Holt (1959) thirds of their maximum. investigated general associations among growth Despite its comparatively lengthy history, the pattern, as shown by the von Bertalanffy growth application and understanding of life-history in- coefficient (K) and asymptotic length (L•), age (a) variants is very much in its infancy. Most of the re- and length at maturity (La), natural mortality (M) search to date has comprised a search for pattern and lifespan (proportional to M-1). In addition to among combinations of life-history metrics. Use- Beverton’s (1992) excellent historical perspective ful as this has been, the challenge now is to formu- of the search for invariants in fish, Charnov (1993) late and test hypotheses that would identify the provides the most comprehensive treatment of processes responsible for the observed patterns. life-history invariants to date. Mangel’s (1996) study on life-history invariants in The practical objective underlying much of the brown trout and Arctic char is an excellent exam- early search for life-history invariants in fish was ple of an attempt to do just that. to identify generalizations which could then be used to estimate natural mortality rate in com- mercially exploited fishes (Beverton 1992). For 7.3 OFFSPRING SIZE AND example, based on an analysis of 175 fish stocks NUMBER STRATEGIES (Pauly 1980; Charnov 1993), the invariant M/K for 7.3.1 Theoretical context teleosts is approximately 1.7, although Charnov (1993) suggests a range of possible values extend- Why do some fish, ranging phylogenetically from ing from 1.6 < M/K < 2.1. If one had an estimate of the sea lamprey (Petromyzon marinus, Cepha- the von Bertalanffy growth coefficient, K, for a laspidomorpha) through the Atlantic sturgeon given stock, a comparatively easy metric to esti- (Acipenser oxyrhynchus, Acipenseriformes) and mate (see Jobling, Chapter 5, this volume), one the Atlantic cod (Gadus morhua, Gadiformes) to could then use the invariant M/K = 1.7 to estimate the sunfish (Mola mola, Tetraodontiformes), natural mortality for the same stock. produce hundreds of thousands, often millions, of Life-history invariants are thus ratios of para- very small eggs (<1.5mm diameter), while other meters and/or variables that have the same units fish, including myxinids, chondrichthyans, many of measure so that the ratios are dimensionless. salmoniforms and mouth-brooding siluriforms Among the most commonly examined life-history produce comparatively few, relatively large eggs invariants in fish are those between (i) mortality (>4mm diameter)? and growth rate, M/K, (ii) length at maturity and The evolutionary implications of the trade-off asymptotic length, La/L•, (iii) age at maturity and between offspring number and offspring size in lifespan, (aM)-1, (iv) age and length at maturity, fish were first considered by the Swedish fish biol- aLa and (v) asymptotic length and growth rate, L•K ogist Gunnar Svärdson (1949). He suggested that (Beverton and Holt 1959; Pauly 1980; Roff 1984; there must be an upper limit to fecundity and that Beverton 1992; Charnov 1993; Vøllestad and this upper limit depends on the influence of egg L’Abée-Lund 1994; Mangel 1996; Jensen 1997; size on offspring survival and parental reproduc- Froese and Binohlan 2000; Reynolds et al. 2001). tive success. Otherwise, he argued, directional In addition to their practical utility, invariants selection – or, as he put it, a tendency to increase have the potential to provide insight into life- egg number every generation – would favour con- history evolution. For example, the invariant M/K tinually increased numbers of eggs per female. implies that fast-growing fish experience higher Svärdson (1949) remarked that, ‘From a theoretical natural mortality rates than slow-growing fish. point of view it is rather easy to conclude that there Life Histories 159

must be a selection pressure for decreasing egg vival curve (the solid curves in Fig. 7.1a). In other numbers, but it is not so extremely evident how words, the optimum corresponds to the egg size at this selection works.’ which the instantaneous rate of gain in fitness per The theoretical underpinning of most research unit increase in offspring size is at its maximum. investigating the adaptive significance of offspring As this instantaneous rate of gain increases, along size and number variability is a graphical model with the slope of the line drawn from the origin, op- proposed by Smith and Fretwell (1974), who asked timal egg size decreases (Winkler and Wallin 1987). how a parent should distribute a fixed amount of Smith and Fretwell’s (1974) model has formed the energy or resources to an indeterminate number of basis of many theoretical treatments of the evolu- young. Optimal egg size is defined graphically tion of egg size, including those examining the by the point on the fitness function at which a effects of parental care (Sargent et al. 1987), food straight line drawn from the origin (the dashed abundance (Hutchings 1991, 1997) and lifetime lines in Fig. 7.1a) is tangential to the offspring sur- reproductive effort (Winkler and Wallin 1987).

(a) (b) 1 High food 1

Medium food High food

Low food Medium food

Low food Juvenile survival Juvenile survival

0 0 Small Large Small Large High Medium Low High/medium/low Egg size Egg size (c) (d) High High Parental fitness Parental fitness High food High food Medium food Medium food Low Low Low food Low food Small Large Small Large High Medium Low High/medium/low Egg size Egg size

Fig. 7.1 Functions relating juvenile survival (a and b) and parental fitness (fecundity ¥ egg-size-specific survival probabilities, the latter for c and d taken from panels a and b, respectively) to egg size in environments differing in food abundance. Solid triangles below the abscissas indicate egg-size optima. (Source: from Hutchings 1997, with kind permission of Kluwer Academic Publishers.) 160 Chapter 7

Strong empirical support for Smith and Fretwell’s 7.3.3 Hypotheses to explain (1974) model was recently obtained from a study variability in strategies of the size and of Atlantic salmon egg size and maternal fitness (Einum and Fleming 2000). number of offspring Various hypotheses have been proposed to explain 7.3.2 How does egg size influence variation in egg size and fecundity within and offspring survival? among species of fish. Many adaptive explanations centre upon proposed selection responses to At first glance, it might seem reasonable to assume species- and age-specific differences in the quality that the larger the egg, the greater an offspring’s of parental care (Sargent et al. 1987) and to seasonal survival probability. Indeed there is considerable (Rijnsdorp and Vingerhoed 1994; Trippel 1998), experimental evidence (Houde 1989; Pepin 1991, population (Kaplan and Cooper 1984; Hutchings 1993; Chambers 1997; Wootton 1998) that larger 1991, 1997) and individual (Jonsson et al. 1996) dif- offspring are produced from larger eggs, have in- ferences in access to food resources. Quinn et al. creased survival during short periods of low food (1995) suggested that among-population variation supply, can be more competitive than smaller off- in sockeye salmon, Oncorhynchus nerka, egg size spring and, by virtue of their faster swimming can be explained as adaptive responses to differ- speeds, may have a lower risk of predation. Field- ences in the size composition of incubation gravel, based data have provided indirect evidence that arguing that the positive association between egg survival in early life may be positively associated size and substrate size may be related to the latter’s with size and individual growth rate at hatching, influence on dissolved oxygen supplies relative to both of which are presumed positive correlates of the surface-to-volume ratio constraints of eggs. egg size (Elliott 1994; Wootton 1998). Gronkjaer The reduction in egg mortality achieved by var- and Schytte (1999) reported that otolith hatch- ious forms of parental care, expressed in the form checks of Baltic cod larvae that survived 20 days of burying of eggs, predator defence, mouthbrood- post-hatch were similar to or larger than the over- ing and egg fanning, is considered a primary selec- all mean, suggesting an association between size at tive factor responsible for the positive association hatch and larval survival. Meekan and Fortier between egg size and amount of parental care (1996) documented both positive and equivocal among species (Sargent et al. 1987; Forsgren et al., evidence that high individual growth rate en- Chapter 10, this volume). Parental care may also hances larval cod survival. Despite these potential offset the mortality costs associated with the benefits to fitness, larger egg sizes may impose longer developmental times of larger eggs. By ex- constraints that negatively influence survival. For tension, the positive association between egg size example, larger eggs can have longer development and maternal size documented within many fish times (Kamler 1992) that prolong a potentially species has been attributed to a greater ability of vulnerable period of life, greater oxygen demands larger females to provide parental care to their (Kamler 1992; Quinn et al. 1995) and may, in young (Sargent et al. 1987). Larger females may be addition to the young that emerge from them, be able to provide greater protection to eggs. How- more visible to predators (Litvak and Leggett 1992; ever, the generality of this hypothesis must be Pepin et al. 1992). Thus, the costs and benefits to tempered by the observation that egg size also offspring of emerging from large and small eggs can increases with female size in fish that provide no be expected to vary with a variety of factors, in- parental care. This is shown by Atlantic cod cluding the habitat into which eggs are released, (Chambers and Waiwood 1996; Kjesbu et al. 1996), such as being buried in substrate, attached to Atlantic herring, Clupea harengus (Hempel and plants/rocks or dispersed freely into the water col- Blaxter 1967), caplin, Mallotus villosus (Chambers umn, parental care, food supply and risk of detec- et al. 1989), and striped bass, Morone saxatilis tion by visual predators. (Zastrow et al. 1989). Life Histories 161

Population differences in average egg size are survival (Fig. 7.1d). Under such circumstances, the often considered a proxy for adaptive variation. evolutionarily stable strategy of investment per But, as previously noted elsewhere (Hutchings offspring would appear to be one of maximizing the 1991; Reznick and Yang 1993), the relationship number of offspring, each approaching the physio- between offspring size and offspring survival logically minimum size, within a brood. must differ among environments, or among popu- Given the within-individual trade-off that lations, for environment- or population-specific must exist between egg size and egg number for a egg-size optima to exist (Fig. 7.1). Hutchings (1991) specific gonadal volume, it is evident that for selec- reported the first such phenotype ¥ environment tion to favour an increase in egg size, the survival interaction on offspring survival in fish. Brook benefits to offspring produced from larger eggs trout survival in the laboratory during the first 50 must exceed the parental fitness cost of producing days following yolk-sac resorption was found to in- fewer eggs (Wootton 1994; Hutchings 1997). Un- crease with egg size, but the effects of egg size and fortunately, explicit recognition of this necessity food abundance on juvenile survival were not addi- is notably rare in many discussions of egg-size opti- tive: decreased food abundance increased mor- ma, particularly in the marine fish literature. The tality among juveniles from the smallest eggs but fecundity cost associated with the production of had no effect on the survival of juveniles produced large eggs must be acknowledged if natural varia- from the largest eggs, a finding similar to that tion in egg size in fish is to be interpreted within an observed for brown trout (Einum and Fleming ecological or evolutionary framework. Demon- 1999). Based on these experimental data, and sup- stration that larger eggs have lower predation or ported by field data on egg size and food abundance starvation mortality than smaller eggs need not in (Hutchings 1997), fitness functions described itself reveal anything about the selective advan- therefrom suggested that low food supply would tage of producing large eggs or of the strength of a favour the production of comparatively few, large year-class composed of large eggs. offspring while high food abundance would favour While the search for adaptive explanations for females that produced many, comparatively small egg-size variability is perhaps more appealing in- offspring (Hutchings 1991). tellectually, researchers need always be cognizant The dependence of environment-specific egg- of the possibility that the observed variation has size optima on the shape of the function relating no adaptive basis. For example, egg size in many offspring survival to egg size is illustrated in Fig. fish is negatively influenced by water tempera- 7.1, where parental fitness (Fig. 7.1c,d) is approxi- ture (Pepin 1991; Kamler 1992; Chambers 1997). mated by the product of egg survival and egg Trippel (1998) also cautions that seasonal declines number, holding gonad volume constant. Two in egg size in batch-spawning marine fish, such as basic functions are considered: the size-dependent Atlantic cod, might reflect a decline in the physio- case, for which offspring survival varies continu- logical condition of the female and a reduced abil- ously with egg size (Fig. 7.1a), and the size- ity to allocate sufficient resources to each egg. It is independent case, for which survival above and also worth considering the high probability that a below a very narrow range of egg sizes is constant considerable amount of the variation in egg size, (Fig. 7.1b). For the former, any factor such as food notably within females, is purely a function of de- supply that is expected to increase offspring velopmental noise or instability (Markow 1994). survival across all egg sizes is predicted to effect a For example, among brook trout in Freshwater reduction in optimal egg size (Fig. 7.1a), thus River, Newfoundland, for which the diameters of favouring females that produce relatively numer- ten randomly chosen eggs were measured for each ous, smaller offspring (Fig. 7.1c). By contrast, if of 114 females (J.A. Hutchings, unpublished data), offspring survival is independent of offspring size, egg volume within females differed by an average optimal egg size is predicted to remain unchanged 23%! Although one could argue that variable egg with changes in a factor that increases offspring sizes within females in this case represents an 162 Chapter 7 adaptive response to unpredictable environmental exist within as well as between alternative heterogeneity (Kaplan and Cooper 1984), the hypo- strategies. thesis that this rather considerable amount of Atlantic salmon males, a case in point, are typi- egg-size variability within females is a function of cally described as exhibiting alternative strategies developmental noise cannot be discounted. (Myers 1986; Thorpe 1986; Hutchings and Myers 1988, 1994; Metcalfe 1998). Following a seaward migration as smolts, anadromous males return to 7.4 ALTERNATIVE LIFE- their natal river and mature comparatively late HISTORY STRATEGIES in life (4–7 years) at a large size (45–90cm). By contrast, mature male parr do not migrate to sea Within populations, significant life-history varia- prior to reproduction, spawning relatively early in tion can exist among individuals of the same sex. A life (1–3 years) at a comparatively small size (<7– common observation is to find some males matur- 15cm), whereafter they may or may not migrate ing relatively early in life, often at a comparatively to sea. Prior to spawning, dominant anadromous small size, attempting to obtain secondary access males defend access to an anadromous female to females by ‘sneaking’ fertilizations in competi- while mature male parr establish what appears to tion with later-maturing larger males who often be a size-based dominance hierarchy immediately have primary access to a female through territorial downstream of the courting anadromous fish or mate-defence behaviours. Taborsky’s (1994) (Jones 1959; Myers and Hutchings 1987b). Mature review identifies 25 families (130 species) in male parr compete with one another and with which alternative reproductive strategies have anadromous males for the opportunity to fertilize been documented, the most common being the eggs (Hutchings and Myers 1988; Thomaz et al. labrids (e.g. Warner 1991), cichlids (e.g. Martin and 1997; Jones and Hutchings 2001, 2002). Taborsky 1997) and salmonids (e.g. Jones 1959; However, it is not simply the gross differences Gross 1984; Thorpe 1986). A considerable litera- in body size and mating behaviour between ture exists on the terminology used to describe anadromous males and mature male parr that re- reproductive strategies and tactics (e.g. Taborsky quire explanation; age at maturity, body size and 1994; Gross 1996), much of which centres on the survival to maturity can differ significantly among degree to which the alternative life histories are of individuals adopting either the parr or anadro- genetic or environmental origin. For simplicity, mous male strategy (Myers et al. 1986; Hutchings and given that the relative contribution of genetic and Myers 1994). Thus, any evaluation of the and environmental factors to the expression of fitness associated with the mature parr and ana- alternative life histories is generally not known, I dromous male phenotypes will be incomplete use the term ‘alternative reproductive strategy’ without explicit consideration of the influence on to refer to the combination of life history and fitness arising from life-history differences within behavioural correlates of distinctive reproductive each of these two strategies. Hutchings and Myers alternatives evident within a sex, within a single (1994) proposed a model to incorporate the exis- population, at one point in time (cf. Henson and tence of multiple age-specific sets of fitness func- Warner 1997). tions within populations of Atlantic salmon. They What precisely constitutes an ‘alternative re- suggested that the fitness of parr and anadromous productive strategy’? The phrase is often used males might best be represented as two multidi- somewhat loosely to describe usually two, occa- mensional fitness surfaces, and that the points of sionally three, sets of reproductive or mating be- intersection of these surfaces specified an evolu- haviours exhibited by members of the same sex tionarily stable continuum of strategy frequencies within a single population. However, by focusing along which the fitness associated with each solely on behaviour, one risks downplaying the ob- strategy would be equal. Approaching the same servation that significant life-history differences, question from a developmental perspective and and thus significant influences on fitness, often from considerations of proximate, rather than ulti- Life Histories 163 mate, influences on fitness, Thorpe et al. (1998) High High have developed a model that also incorporates fitness surfaces to explain life-history variation within salmonids.

7.4.1 Maintenance of alternative reproductive strategies There are several means by which alternative maturation phenotypes can be maintained within Fitness of parr strategy ( ) populations. As with many characters, the origins Low Low Fitness of andadromous strategy ( ) of alternative phenotypes may be purely genetic, purely environmental, or some combination 0 25 50 75 100 thereof. If alternative strategies have any genetic Incidence of parr strategy (%) basis, then the fitness associated with each maturation genotype must, on average, be approx- Fig. 7.2 Graphical representation of negative imately equal, otherwise the less fit genotype frequency-dependent selection in male Atlantic salmon would be continually selected against and eventu- life histories. As the incidence of either the mature male parr or anadromous male strategy increases within a ally disappear from the population. Under such population, the average fitness of individuals adopting circumstances, the frequencies of the strategy- that strategy declines. The equilibrium frequencies of specific genotypes within populations are said to the two strategies, at which individual fitnesses are be evolutionarily stable if such populations cannot equal, is identified by the point of intersection of the two be invaded by genotypes adopting other strategies fitness functions. (Maynard Smith 1982). Negative frequency-dependent selection has been hypothesized to be the primary means by some species may be under complete genetic con- which alternative mating genotypes are main- trol exists for the pygmy swordtail, Xiphophorus tained within populations (e.g. Gross 1984; Myers nigrensis. Zimmerer and Kallman (1989) found 1986; Partridge 1988). As the frequency of a given that the four mating phenotypes in this species, strategy increases within a population, increased distinguishable by body size and behaviour, appear competition among individuals adopting that to be controlled by genetic variation at a Y-linked strategy would lead to reduced average fitness locus. Breeding experiments suggest that the male among those individuals (see the dashed curve in parr and anadromous male phenotypes in Atlantic Fig. 7.2). In contrast, the average fitness of individ- salmon have a heritable basis (Naevdal et al. 1976; uals adopting the alternative strategy would in- Thorpe et al. 1983; Glebe and Saunders 1986; crease because of reduced competition (solid curve Herbinger 1987). Also, the relatively larger testis in Fig. 7.2), resulting in a shift in the incidence of size in the subordinate of two strategies also pro- strategies towards the rarer of the two. In theory, vides compelling evidence of a genetic basis to then, the frequency of alternative strategies alternative maturation phenotypes in corkwing within a population eventually achieves an equi- wrasse, Symphodus melops (Uglem et al. 2000), librium at which the average fitness of individ- Atlantic salmon (Gage et al. 1995) and plainfin uals adopting each strategy is equal (Fig. 7.2). midshipman, Porichthys notatus (Brantley and Despite its prevalence in theory, evidence of a Bass 1994). genetic basis for alternative mating phenotypes is Within-population variation in reproductive limited, although it is not clear if this is due to very strategies may not be under direct genetic control, low or negligible levels of heritability or to the reflecting instead a breadth of mating behaviours rarity of research devoted to this question. Evi- and ages at maturity that are purely a consequence dence that alternative reproductive strategies in of stochastic variability in the environment. Thus, 164 Chapter 7 individuals attempting to sneak fertilizations (Myers et al. 1986; Thorpe 1986; Metcalfe 1998) in competition with behaviourally dominant and there is evidence that male parr maturity has individuals may be engaging in this behaviour be- a genetic basis (Naevdal et al. 1976; Thorpe et al. cause of environmentally predicated influences on 1983; Glebe and Saunders 1986; Herbinger 1987). phenotypic attributes that have reduced the prob- To incorporate both the environmental and ability of obtaining a mate. Examples would be genetic determinants of alternative reproductive comparatively smaller body size or poorer body strategies in this species, parr maturation can condition. Under such circumstances, one would be modelled as a threshold trait (Myers and predict that the fitness of individuals adopting Hutchings 1986; Falconer 1989) such that adop- such suboptimal maturation phenotypes, which tion of either the parr or the anadromous strategy may be fixed for a breeding season or for life, would depends on whether an individual’s growth rate in be less than that of males adopting alternative early life exceeds that specified by a genetically behaviours. determined growth-rate threshold (Myers and On the other hand, individuals might adopt dif- Hutchings 1986; Thorpe 1986; Hazel et al. 1990; ferent maturation phenotypes as they age or as en- Hutchings and Myers 1994; Thorpe et al. 1998). vironmental circumstances change within a single Theoretical representations of such growth- breeding season, possibly reflecting a condition- rate thresholds for male parr maturity are pre- or status-dependent behavioural or life-history sented in Fig. 7.3. Sigmoid curves describe how response. For example, bluehead wrasse, Thalas- the probability of parr maturity might increase soma bifasciatum, adopt non-territorial with parr growth rate, approximated here by behaviours when they are young, becoming size-at-age, among four populations. The lengths- territorial as they grow older and larger (Hoffman at-age at which 50% of the males are expected to et al. 1985). Another reef fish, the peacock wrasse, adopt the parr strategy can be identified as parr Symphodus tinca, will facultatively switch be- maturation thresholds, which would be averaged tween spawning in its own territory and within among all male parr within a population. those of other males (van den Berghe 1990). Thorpe Population differences in demography would be et al.’s (1998) stochastic dynamic modelling of expected to generate population differences in parr male Atlantic salmon life histories explores how proximate changes to body condition might influence the adoption of the parr or anadromous 1 male strategy in Atlantic salmon. Forsgren et al. (Chapter 10, this volume) deal further with the 0.8 behavioural aspects of alternative reproductive 0.6 strategies. A third, and not necessarily mutually exclu- 0.4 sive, means by which alternative reproductive strategies can be maintained within populations is 0.2 through some form of adaptive phenotypic plastic- Incidence of male parr maturity ity (Warner 1991). One means of incorporating 0 both environmental and genetic influences on the expression of alternative reproductive strategies 60 65 70 75 80 85 90 was proposed by Hutchings and Myers (1994) for Parr growth rate (length at age 1+ years) Atlantic salmon. For example, the expression of Fig. 7.3 Growth-rate thresholds for male parr maturity the mature male parr strategy in Atlantic salmon among four hypothetical populations of Atlantic appears to be influenced by both genetic and envi- salmon. The solid triangles represent the lengths-at-age ronmental factors; the fastest growing males in a at which 50% of the male parr in each population would population are those most likely to mature as parr be expected to adopt the parr reproductive strategy. Life Histories 165 growth-rate thresholds. Evidence of such sigmoid associated with increased variation in repro- relationships exists for male parr in Newfound- ductive success which would reduce Ne. Indeed, land’s Little Codroy River (Myers et al. 1986; genetic differences among small Atlantic salmon Hutchings and Myers 1994), the only natural populations may be primarily attributable to population for which individual parr maturity mature male parr (M.W. Jones, unpublished data). and growth-rate data have been published. Here it These hypotheses merit study. is the maturation response to environmental stimuli, be they behavioural or environmental, that would be under selection. Note that the actual strategies adopted need not have equal 7.5 EFFECTS OF FISHING fitness, but selection for the reaction norm under- ON LIFE HISTORY lying the environmental-, condition- or status- mediated response might result in negative Phenotypic and genetic life-history responses to frequency-dependent selection for alternative fishing can be categorized as follows. growth-rate thresholds within populations 1 A comparatively rapid (<1 generation) pheno- (Hutchings and Myers 1994). typic response effected by increased individual growth rates concomitant with reductions in fish 7.4.2 Management implications density (Policansky 1993). 2 A comparatively rapid (<1 generation) pheno- From a management perspective, the incidence of typic change effected by plastic responses along male parr maturity has at least three implications. reaction norms for life-history traits such as age Firstly, the higher the incidence of male parr matu- at maturity and reproductive effort (McKenzie rity, the lower the production of anadromous male et al. 1983; Hutchings 1993b, 1997; Nelson 1993). salmon. This can be attributed to the higher mor- 3 A comparatively rapid (<1 generation) genetic tality experienced by mature male parr in fresh change at the population level caused by extre- water relative to their immature counterparts mely high mortality rates that differentially (Myers 1984; Hutchings and Jones 1998). For ex- and rapidly reduce the frequency of some life- ample, the increased mortality and delayed age at history genotypes, e.g. late-maturing genotypes, smolting experienced by mature male parr may be relative to others, e.g. early-maturing geno- responsible for the loss of 60% of the anadromous types (Hutchings 1999). male salmon production in some populations 4 A comparatively slow (10–30 generations) (Myers 1984). A second consideration is that genetic response effected by selection against spawning contributions by mature male parr are life-history genotypes, or against norms of reac- not explicitly incorporated into stock assessments tion, whose fitness in the presence of fishing is less of Atlantic salmon populations. This exclusion than that of other genotypes, or other norms of may affect the interpretation of the spawning es- reaction, in the population (Law and Grey 1989; capement of males required to maintain a viable Hutchings 1997; Hendry and Kinnison 1999). population, perhaps most notably for populations Although categories 1 and 2 are both phenotyp- in which the incidence of male parr maturity is ic responses, their distinction is a useful one be- quite high. Thirdly, from a conservation per- cause the latter makes explicit the hypothesis that spective, it is not clear how the incidence of parr reaction norms for life-history characters exist and maturity influences effective population size, or that the pattern of phenotypic responses to envi-

Ne (Jones and Hutchings 2001, 2002). Populations ronmental change can therefore have an adap- with low escapement of anadromous males may tive basis (Haugen 2000; Haugen and Vøllestad have a comparably high Ne because of the genetic 2000). The primary distinction between the two contributions of mature male parr. Alternatively, categories of genetic response is that short-term increased incidence of male parr maturity may be genetic changes in gene frequencies (category 3) 166 Chapter 7 do not depend on the level of heritability of a life- questions remain, of course, whether such plastic history trait, although they are assumed to be heri- changes in age at maturity are adaptive and table. By contrast, long-term genetic responses by whether the affected population’s growth rate, and a trait to selection, R, are influenced both by that its associated risk of overfishing following a re- trait’s heritability, h2, and by the difference in the duced age at maturity, will be negatively affected average value of that trait among reproductive by such changes. Beverton et al. (1994), for exam- individuals relative to the average in the popula- ple, estimated that early-maturing (6 and 7 years tion as a whole, or in other words the selection old) northeast Arctic cod experience considerably differential, S (Falconer 1989), such that R = h2S. higher natural mortality rates (0.25 and 0.17 re- The predominant changes to life history associ- spectively) than those that delay maturity to age ated with fishing are reduced age and size at matu- 8 and later (M = 0.15). Thus, a reduction in age at rity, the latter often being a simple consequence of maturity, unaccompanied by any fitness advan- the former, although increases in both characters tage associated with such a response (see category might occur under certain circumstances (Heino 2 above), can have serious consequences to a popu- 1998; Rochet 1998). The rapidity with which many lation’s growth rate and persistence. of these changes has occurred is consistent with Based upon age-specific survival, fecundity and the hypothesis of a phenotypically plastic response growth-rate data for several unexploited brook to exploitation. In theory, reductions in density trout populations, Hutchings (1993b, 1997) pre- effected by fishing should lead to reduced com- dicted how optimal norms of reaction for age at petition for resources, resulting in an increase in maturity and reproductive effort might change individual growth rate and possibly body condi- under increasing levels of exploitation (Fig. 7.4). tion. Given the widely documented negative The magnitude of change will depend on the shape association between individual growth rate and of the reaction norm, on the magnitude of change age at maturity in fish (e.g. Alm 1959; Roff 1992; in individual growth rate and on the heritability for Hutchings 1993a), a comparatively rapid decline in the shape of the reaction norm. Although these age at maturity can be explained as a plastic re- changes would be reversible in the short term, sponse to increases in individual growth. Such as- over perhaps less than ten generations, persis- sociations have been documented in exploited tently strong selection effected by fishing might populations of marine fish such as Atlantic cod select against the shape of plastic reaction norms (Chen and Mello 1999), yellowtail flounder (Walsh under zero or low fishing mortality, as shown by and Morgan 1999) and North Sea plaice (Rijnsdorp the dashed reaction norms in Fig. 7.4, favouring 1993). other, possibly non-plastic, reaction norms that Increases to individual growth rate and reduc- are optimal under high fishing pressure. These are tions in adult survival are also predicted to in- shown by the solid reaction norms in Fig. 7.4. crease reproductive effort. Although there has Thus, although initial changes to life history ef- been considerably less attention directed to such fected by exploitation may be plastic and therefore changes, there is evidence that temporal changes reversible, persistently high selection intensities in size-specific fecundity in the orange roughy, may effect genetic changes in the shapes of reac- Hoplostethus atlanticus, may reflect a life-history tion norms and the loss of plasticity for the trait(s) response in reproductive effort to fishing. Between in question within the exploited population (see 1987 and 1992, when the commercially exploited Haugen 2000 for a potential example in grayling, orange roughy stock off east Tasmania was re- Thymallus thymallus). duced by 50%, individual fecundity increased While some short-term changes in age at matu- 20% on average (Koslow et al. 1995). Law (1979) rity appear to be linked to increases in individual reported a 60% increase in the fecundity of 3- growth rate, and can potentially be explained as year-old northern pike, Esox lucius, 12 years after phenotypically plastic responses to fishing, others an experimental harvest in Windermere, UK. The are not. One example is that of the northern stock Life Histories 167

Concern that selection induced by fishing gear (a) Unfished Late Low exploitation rate might effect genetic responses in fish populations High exploitation rate has been discussed at least since the 1950s (Miller 1957). Long-term changes in size at maturity, for example, have been interpreted as selective re- sponses to the size-selectivity of fishing gear, no- tably gill-nets, as seen in lake whitefish (Handford Age at maturity et al. 1977), pink salmon (Oncorhynchus kisutch) Early and chinook salmon (O. tshawytscha; Ricker 1981), and Atlantic salmon (Bielak and Power Low High 1986). Evidence that long-term changes in age at Growth rate/body condition/(1/density) maturity may represent evolutionary responses (b) to exploitation has also been forthcoming in High studies of cod (Rowell 1993; Law 2000) and plaice (Rijnsdorp 1993) in the North Sea. As expressed succinctly by Rijnsdorp (1993), fisheries are large-scale experiments on life- history evolution. The potential for fishing to effect significant evolutionary change within a Reproductive effort Unfished Low population is no different from that of any other Low exploitation rate form of predator-induced mortality that differen- High exploitation rate tially affects the survival of individuals of different Low High ages and sizes. The question is not whether fishing Growth rate/body condition/(1/density) represents a primary selective pressure effecting evolutionary change in exploited fish populations Fig. 7.4 Hypothetical norms of reaction for age at – clearly it must. The important questions surely maturity (a) and reproductive effort (b) for a fish concern the type of life-history responses (see population in an unfished state and at low and high categories 1–4 above), the reversibility of the levels of exploitation. The abscissa reflects an environmental gradient that may have influences on responses and the consequences of the responses individual growth rate, body condition and density. to population growth rate and likelihood of population recovery (Hutchings 2000; Law 2000, 2001). of Atlantic cod extending from southeastern Labrador to the northern half of Newfoundland’s 7.6 CONCLUSIONS Grand Bank. Between the mid-1980s and the mid- 1990s, female median age at maturity declined Life histories ultimately determine an individual’s by more than 1 year, a reduction of approximately fitness, a population’s persistence and growth rate 17% (Lilly et al. 1998). However, these changes at low abundance, and a commercially exploited were not associated with either increases in indi- stock’s ability to sustain exploitation. The study of vidual growth rate or increases in body condition life histories is at the heart of research addressing (Lilly et al. 1998). Hutchings (1999) suggested the evolutionary ecology, conservation and ex- that the most parsimonious explanation for these ploitation of fish. There is a need for studies changes in age at maturity was an extremely rapid that focus on the underlying basis for life-history differential reduction of late-maturing genotypes change in fish. Are life-history responses to envi- by severe overfishing, relative to that experienced ronmental change generally effected by changes in by early-maturing genotypes. gene frequencies or by phenotypic modification 168 Chapter 7 along norms of reaction? Compared with theoreti- REFERENCES cal interest in the former, there has been a lack of research addressing genotype-by-environment in- Adams, B.K. (1999) Variation in genetic structure and life teractions in fish life-history responses. Such work histories among populations of brook char, Salvelinus might attempt, for example, to describe norms fontinalis, from insular Newfoundland, Canada. MSc of reaction for life-history traits, to quantify thesis, Dalhousie University, Halifax, Nova Scotia, the heritability of reaction norms (Haugen & Canada. Vøllestad 2000), to assess the degree to which Adlerstein, S.A. and Dorn, M.W. (1998) The effect of the shapes of reaction norms can be modified by Kudoa paniformis infection on the reproductive effort Canadian Journal of Zoology selection, and to provide empirical tests for the of female Pacific hake. 76, 2285–9. existence of genetically determined, environmen- Alm, G. (1959) Connection between maturity, size and tally triggered threshold characters. age in fishes. Report of the Institute of Freshwater Tests of life-history theory are hampered by a Research, Drottningholm 40, 5–145. lack of empirical data on survival and reproductive Balshine-Earn, S. (1995) The costs of parental care in costs in wild populations. Despite the difficulty of Galilee St. Peter’s fish, Sarotherodon galilaeus. quantifying these parameters, their absence leaves Animal Behaviour 50, 1–7. Belk, M.C. (1995) Variation in growth and age at maturity purported demonstrations of adaptive life-history in bluegill sunfish: genetic or environmental effects. responses to selection severely wanting. Even Journal of Fish Biology 47, 237–47. when data on natural mortality (M) are presented, Bell, G. (1980) The costs of reproduction and their conse- there is rarely any attempt on the part of the re- quences. American Naturalist 116, 45–76. searcher(s) to distinguish that part of the mortality Beverton, R.J.H. (1992) Patterns of reproductive strategy attributable to external factors from the mortality parameters in some marine teleost fishes. Journal of consequences of a particular genotype’s reproduc- Fish Biology 41 (Suppl. B), 137–60. Beverton, R.J.H. and Holt, S.J. (1959) A review of the tive decisions regarding its age and effort at matu- lifespans and mortality rates of fish in nature, and rity. In other words, the use of M as a metric of the their relation to growth and other physiological degree to which mortality favours a particular life- characteristics. CIBA Foundation Colloquia on history decision, such as age at maturity, results in Ageing 54, 142–80. a circular argument given that the life-history con- Beverton, R.J.H., Hylen, A. and Ostvedt, O.J. (1994) sequences of that genotype’s decision as to when to Growth, maturation, and longevity of maturation co- ICES Journal of Marine mature are already embedded in M. The same is horts of Northeast Arctic cod. Science Symposium 198, 482–501. true when using K, the von Bertalanffy growth co- Bielak, A.T. and Power, G. (1986) Changes in mean efficient, as a metric of the degree to which growth weight, sea-age composition, and catch-per-unit-effort rate favours a particular life-history decision: K of Atlantic (Salmo salar) angled in the Godbout River, already incorporates the growth costs of repro- Quebec, 1859–1983. Canadian Journal of Fisheries duction effected by a genotype’s age and effort and Aquatic Sciences 43, 281–7. at maturity. Bond, C.E. (1996) Biology of Fishes. Orlando, FL: Saunders College Publishing. Perhaps the ultimate value in studying fish life Brantley, R.K. and Bass, A.H. (1994) Alternative male histories lies in the hope of eventually being able to spawning tactics and acoustic signals in the plain- predict how short- and long-term anthropogenic, fin midshipman fish Porichthys notatus Girard biological and physical environmental perturba- (Teleostei, Batrachoididae). Ethology 96, 213–32. tions influence age-specific rates of survival and Chambers, R.C. (1997) Environmental influences on fecundity, rates of increase, population persistence egg and propagule sizes in marine fishes. In: R.C. and community composition. Based on the theo- Chambers and E.A. Trippel (eds) Early Life History and Recruitment in Fish Populations. London: retical and empirical strengths of fish life-history Chapman & Hall, pp. 63–102. research over the past 50 years, there is room for Chambers, R.C. and Waiwood, K.G. (1996) Maternal considerable optimism that such an objective can and seasonal differences in egg sizes and spawning be achieved. characteristics of captive Atlantic cod, Gadus morhua. Life Histories 169

Canadian Journal of Fisheries and Aquatic Sciences Fleming, I.A. (1996) Reproductive strategies of Atlantic 53, 1986–2003. salmon: ecology and evolution. Reviews in Fish Biol- Chambers, R.C., Leggett, W.C. and Brown, J.A. (1989) Egg ogy and Fisheries 6, 379–416. size, female effects, and the correlations between early Foote, C.J. (1988) Mate choice dependent on male size in life history traits of capelin (Mallotus villosus): an salmon. Behaviour 106, 63–80. appraisal at the individual level. Fishery Bulletin 87, Fox, M.G. (1994) Growth, density, and interspecific 515–23. influences on pumpkinseed sunfish life-histories. Chapman, T., Hutchings, J. and Partridge, L. (1993) No Ecology 75, 1157–71. reduction in the cost of mating for Drosophila Fox, M.G. and Keast, A. (1991) Effect of overwinter mor- melanogaster females mating with spermless males. tality on reproductive life history characteristics of Proceedings of the Royal Society of London Series B pumpkinseed (Lepomis gibbosus) populations. 253, 211–17. Canadian Journal of Fisheries and Aquatic Sciences Charlesworth, B. (1994) Evolution in Age-structured 48, 1791–9. Populations. Cambridge: Cambridge University Press. Froese, R. and Binohlan, C. (2000) Empirical relation- Charlesworth, B. and León, J.A. (1976) The relation of ships to estimate asymptotic length, length at first reproductive effort to age. American Naturalist 110, maturity and length at maximum yield per recruit 449–59. in fishes, with a simple method to evaluate length Charnov, E.L. (1993) Life History Invariants: Some frequency data. Journal of Fish Biology 56, 758–73. Explorations of Symmetry in Evolutionary Ecology. Gadgil, M. and Bossert, W. (1970) Life historical conse- Oxford: Oxford University Press. quences of natural selection. American Naturalist Chen, Y. and Mello, L.G.S. (1999) Growth and matura- 104, 1–24. tion of cod (Gadus morhua) of different year classes Gage, M.G.J., Stockley, P. and Parker, G.A. (1995) Effects in the Northwest Atlantic, NAFO subdivision 3Ps. of alternative mating strategies on characteristics Fisheries Research 42, 87–101. of sperm production in the Atlantic salmon (Salmo Cole, L.C. (1954) The population consequences of life salar): theoretical and empirical investigations. Philo- history phenomena. Quarterly Review of Biology 29, sophical Transactions of the Royal Society of London 103–37. Series B 350, 391–9. Dufresne, F., FitzGerald, G.J. and Lachance, S. (1990) Gibson, R.J., Williams, D.D., McGowan, C. and Age and size-related differences in reproductive suc- Davidson, W.S. (1996) The ecology of dwarf fluvial cess and reproductive costs in threespine stickleback Atlantic salmon, Salmo salar L., cohabiting with (Gasterosteus aculeatus). Behavioural Ecology 1, brook trout, Salvelinus fontinalis (Mitchill), in south- 140–7. eastern Newfoundland, Canada. Polskie Archiwum Dutil, J.-D. (1986) Energetic constraints and spawning Hydrobiologii 43: 145–66. interval in the anadromous Arctic charr (Salvelinus Giloolly, J.F. and Baylis, J.R. (1999) Reproductive success alpinus). Copeia 1986, 945–55. and the energetic cost of parental care in male small- Einum, S. and Fleming, I.A. (1999) Maternal effects of egg mouth bass. Journal of Fish Biology 54, 573–84. size in brown trout (Salmo trutta): norms of reaction Glebe, B.D. and Saunders, R.L. (1986) Genetic factors in to environmental quality. Proceedings of the Royal sexual maturity of cultured Atlantic Salmon (Salmo Society of London Series B 266, 2095–100. salar) parr and adults reared in sea cages. In: D. Einum, S. and Fleming, I.A. (2000) Highly fecund Meerburg (ed.) Salmonid Age at Maturity. Ottawa: mothers sacrifice offspring survival to maximize Department of Fisheries and Oceans, Canadian Spe- fitness. Nature 405, 565–7. cial Publication of Fisheries and Aquatic Sciences 89, Elliott, M. (1994) Quantitative Ecology and the Brown pp. 24–9. Trout. Oxford: Oxford University Press. Godø, O.R. and Haug, T. (1999) Growth rate and sexual Falconer, D.S. (1989) Introduction to Quantitative maturity in cod (Gadus morhua) and Atlantic halibut Genetics. London: Longman. (Hippolglossus hippoglossus). Journal of Northwest Ferguson, M.M., Danzmann, R.G. and Hutchings, J.A. Atlantic Fishery Science 25, 115–23. (1991) Incongruent estimates of population dif- Grantner, A. and Taborsky, M. (1998) The metabolic ferentiation among brook charr from Cape Race, rates associated with resting, and with the perfor- Newfoundland, Canada, based upon allozyme and mance of agonistic, submissive and digging behaviours mitochondrial DNA variation. Journal of Fish Biology in the cichlid fish Neolamprologus pulcher (Pisces: 39 (Suppl. A), 79–85. Cichlidae). Journal of Comparative Physiology B 168, Fisher, R.A. (1930) The Genetical Theory of Natural 427–33. Selection. Oxford: Oxford University Press. Gronkjaer, P. and Schytte, M. (1999) Non-random mor- 170 Chapter 7

tality of Baltic cod larvae inferred from otolith hatch- The costs of changing sex and the ontogeny of males check sizes. Marine Ecology Progress Series 181, 53–9. under contest competition for mates. Evolution 39, Gross, M.R. (1984) Sunfish, salmon, and the evolution 915–27. of alternative reproductive strategies and tactics in Houde, A.E. and Endler, J.A. (1990) Correlated evolution fishes. In: G.W. Potts and R.J. Wootton (eds) Fish Repro- of female mating preferences and male patterns in the duction: Strategies and Tactics. London: Academic guppy, Poecilia reticulata. Science 248, 1405–8. Press, pp. 55–75. Houde, E.D. (1989) Comparative growth, mortality, and Gross, M.R. (1996) Alternative reproductive strategies energetics of marine fish larvae: temperature and im- and tactics: diversity within sexes. Trends in Ecology plied latitudinal effects. Fishery Bulletin 87, 471–95. and Evolution 11, 92–8. Huitfeldt-Kaas, H. (1946) Tribes of salmon in Norway. Handford, P., Bell, G. and Reimchen, T. (1977) A gillnet Magasin for Naturvidenskapene B85, 115–59. fishery considered as an experiment in artificial Hutchings, J.A. (1991) Fitness consequences of variation selection. Journal of the Fisheries Research Board of in egg size and food abundance in brook trout, Salveli- Canada 34, 954–61. nus fontinalis. Evolution 45, 1162–8. Haugen, T.O. (2000) Growth and survival effects on Hutchings, J.A. (1993a) Adaptive life histories effected maturation pattern in populations of grayling with by age-specific survival and growth rate. Ecology 74, recent common ancestors. Oikos 90, 107–18. 673–84. Haugen, T.O. and Vøllestad, L.A. (2000) Population dif- Hutchings, J.A. (1993b) Reaction norms for reproductive ferences in early life-history traits in grayling. Journal traits in brook trout and their influence on life history of Evolutionary Biology 13, 897–905. evolution effected by size-selective harvesting. In: T.K. Hazel, W.N., Smock, R. and Johnson, M.D. (1990) A poly- Stokes, J.M. McGlade and R. Law (eds) The Exploita- genic model for the evolution and maintenance of con- tion of Evolving Resources. Berlin: Springer-Verlag, ditional strategies. Proceedings of the Royal Society of pp. 107–25. London Series B 242, 181–7. Hutchings, J.A. (1994) Age- and size-specific costs of re- Heino, M. (1998) Management of evolving fish stocks. production within and among populations of brook Canadian Journal of Fisheries and Aquatic Sciences trout, Salvelinus fontinalis. Oikos 70, 12–20. 55, 1971–82. Hutchings, J.A. (1996) Adaptive phenotypic plasticity Helfman, G.S., Collette, B.B. and Facey, D.E. (1997) The in brook trout, Salvelinus fontinalis, life histories. Diversity of Fishes. Oxford: Blackwell Science. Écoscience 3, 25–32. Hempel, G. and Blaxter, G.H.S. (1967) Egg weight in Hutchings, J.A. (1997) Life history responses to environ- Atlantic herring (Clupea harengus L.). Journal du mental variability in early life. In: R.C. Chambers and Conseil, Conseil International pour l’Exploration E.A. Trippel (eds) Early Life History and Recruitment de la Mer 31, 170–95. in Fish Populations. London: Chapman & Hall, pp. Hendry, A.P. and Kinnison, M.T. (1999) The pace of 139–68. modern life: measuring rates of contemporary micro- Hutchings, J.A. (1999) The influence of growth and sur- evolution. Evolution 53, 1637–53. vival costs of reproduction on Atlantic cod, Gadus Henson, S.A. and Warner, R.R. (1997) Male and female morhua, population growth rate. Canadian Journal of alternative reproductive behaviors in fishes: a new Fisheries and Aquatic Sciences 56, 1612–23. approach using intersexual dynamics. Annual Review Hutchings, J.A. (2000) Collapse and recovery of marine of Ecology and Systematics 28, 571–92. fishes. Nature 406, 882–5. Herbinger, C.M. (1987) A study of Atlantic Salmon Hutchings, J.A. and Jones, M.E.B. (1998) Life history (Salmo salar) maturation using individually identified variation and growth rate thresholds for maturity in fish. PhD thesis, Dalhousie University, Halifax, Nova Atlantic salmon, Salmo salar. Canadian Journal of Scotia. Fisheries and Aquatic Sciences 55 (Suppl. 1), 22–47. Hinch, S.G. and Collins, N.C. (1991) Importance of diur- Hutchings, J.A. and Myers, R.A. (1987) Escalation of an nal and nocturnal nest defence in the energy budget asymmetric contest: mortality resulting from mate of male smallmouth bass: insights from direct video competition in Atlantic salmon, Salmo salar. observations. Transactions of the American Fisheries Canadian Journal of Zoology 65, 766–8. Society 120, 657–63. Hutchings, J.A. and Myers, R.A. (1988) Mating success of Hirshfield, M.F. and Tinkle, D.W. (1975) Natural selec- alternative maturation phenotypes in male Atlantic tion and the evolution of reproductive effort. Proceed- salmon, Salmo salar. Oecologia 75, 169–74. ings of the National Academy of Sciences of the USA Hutchings, J.A. and Myers, R.A. (1994) The evolution 72, 2227–31. of alternative mating strategies in variable environ- Hoffman, S.G., Schildhauer, M.P. and Warner, R.R. (1985) ments. Evolutionary Ecology 8, 256–68. Life Histories 171

Hutchings, J.A., Pickle, A., McGregor-Shaw, C.R. and Law, R. (1979) Optimal life histories under age-specific Poirier, L. (1999) Influence of sex, body size, and repro- predation. American Naturalist 114, 399–417. duction on overwinter lipid depletion in brook trout. Law, R. (2000) Fishing, selection, and phenotypic evolu- Journal of Fish Biology 55, 1020–8. tion. ICES Journal of Marine Science 57, 659–68. Jansen, W.A. (1996) Plasticity in maturity and fecundity Law, R. (2001) Phenotypic and genetic changes due to ex- of yellow perch, Perca flavescens (Mitchill): com- ploitation. In: J.D. Reynolds, G.M. Mace, K.H. Redford parisons of stunted and normal-growing populations. and J.G. Robinson (eds) Conservation of Exploited Annales Zoologici Fennici 33, 403–15. Species. Cambridge: Cambridge University Press, pp. Jensen, A.L. (1997) Origin of the relation between K and 323–42.

Linf and synthesis of relations among life history pa- Law, R. and Grey, D.R. (1989) Evolution of yields from rameters. Canadian Journal of Fisheries and Aquatic populations with age-specific cropping. Evolutionary Sciences 54, 987–9. Ecology 3, 343–59. Jones, J.W. (1959) The Salmon. London: Collins. Leggett, W.C. and Carscadden, J.E. (1978) Latitudinal Jones, M.W. and Hutchings, J.A. (2001) The influence of variation in reproductive characteristics of American male parr body size and mate competition on fertiliza- shad (Alosa sapidissima): evidence for population tion success and effective population size in Atlantic specific life history strategies in fish. Journal of the salmon. Heredity 86, 675–84. Fisheries Research Board of Canada 35, 1469–78. Jones, M.W. and Hutchings, J.A. (2002) Individual Leggett, W.C. and Power, G. (1969) Differences between variation in Atlantic salmon fertilization success: im- two populations of landlocked Atlantic salmon (Salmo plications for effective population size. Ecological salar) in Newfoundland. Journal of the Fisheries Applications 12, 184–93. Research Board of Canada 26, 1585–96. Jonsson, N., Jonsson, B. and Hansen, L.P. (1991) Energetic Lilly, G.R., Shelton, P.A., Brattey, J. et al. (1998) An as- cost of spawning in male and female Atlantic salmon sessment of the cod stock in NAFO Divisions 2J + 3KL. (Salmo salar L.). Journal of Fish Biology 39, 739–44. Canadian Stock Assessment Secretariat Research Jonsson, N., Jonsson, B. and Fleming, I.A. (1996) Does Document 98/15. Ottawa: Department of Fisheries early growth cause a phenotypically plastic response and Oceans. in egg production of Atlantic salmon? Functional Lindström, K. (1998) Energetic constraints on mating Ecology 10, 89–96. performance in the sand goby. Behavioural Ecology 9, Jørgensen, E.H., Johansen, S.J.S. and Jobling, M. (1997) 297–300. Seasonal patterns of growth, lipid deposition and lipid Litvak, M.K. and Leggett, W.C. (1992) Age and size- depletion in anadromous Arctic charr. Journal of Fish selective predation on larval fishes: the bigger-is- Biology 51, 312–26. better hypothesis revisited. Marine Ecology Progress Kamler, E. (1992) Early Life History of Fish: An Energet- Series 81, 13–24. ics Approach. London: Chapman & Hall. Mackereth, R.W., Noakes, D.L.G. and Ridgway, M.S. Kaplan, R.H. and Cooper, W.S. (1984) The evolution of de- (1999) Size-based variation in somatic energy reserves velopmental plasticity in reproductive characteristics: and parental expenditure by male smallmouth bass, an application of the ‘adaptive coin flipping’ principle. Micropterus dolomieu. Environmental Biology of American Naturalist 123, 393–410. Fishes 56, 263–75. Kjesbu, O.S., Solemdal, P., Bratland, P. and Fonn, M. Mangel, M. (1996) Life history invariants, age at ma- (1996) Variation in annual egg production in individual turity, and the ferox trout. Evolutionary Ecology captive Atlantic cod (Gadus morhua). Canadian Jour- 10, 249–63. nal of Fisheries and Aquatic Sciences 53, 610–20. Markow, T.A. (1994) Developmental Instability: Its Koslow, J.A., Bell J., Virtue, P. and Smith, D.C. (1995) Origins and Evolutionary Implications. Dordrecht: Fecundity and its variability in orange roughy: effects Kluwer Academic Publishers. of population density, condition, egg size, and senes- Martin, E. and Taborsky, M. (1997) Alternative male cence. Journal of Fish Biology 47, 1063–80. mating tactics in a cichlid, Pelvicachromis pulher: a Kozlowski, J. and Uchmanski, J. (1987) Optimal in- comparison of reproductive effort and success. Beha- dividual growth and reproduction in perennial species vioral Ecology and Sociobiology 41, 311–19. with indeterminate growth. Evolutionary Ecology 1, Martini, F., Heiser, J.B. and Lesser, M.P. (1997) A popula- 231–44. tion profile for Atlantic hagfish, Myxine glutinosa (L.), Lande, R. (1993) Risks of population extinction from in the Gulf of Maine. Part I: morphometrics and repro- demographic and environmental stochasticity, and ductive state. Fishery Bulletin 95, 311–20. random catastrophes. American Naturalist 142, 911– Maynard Smith, J. (1982) Evolution and the Theory of 27. Games. Cambridge: Cambridge University Press. 172 Chapter 7

McKenzie, W.D., Crews, D., Kallman, K.D., Policansky, tions. Canadian Journal of Fisheries and Aquatic D. and Sohn, J.J. (1983) Age, weight and the genetics of Sciences 55, 1034–42. sexual maturation in the playfish, Xiphophorus macu- Nelson, K. (1993) Individual variation in acquisition/ latus. Copeia 1983, 770–4. allocation reaction norms. In: T.K. Stokes, J.M. Meekan, M.G. and Fortier, L. (1996) Selection for fast McGlade and R. Law (eds) The Exploitation of Evolv- growth during the larval life of Atlantic cod, Gadus ing Resources. Berlin: Springer-Verlag, pp. 91–106. morhua, on the Scotian Shelf. Marine Ecology Progress Pampoulie, C., Morand, S., Lambert, A., Rosecchi, E., Series 137, 25–37. Bouchereau, J.L. and Crivelli, A.J. (1999) Influence Metcalfe, N.B. (1998) The interaction between behaviour of the trematode Aphalloides coelomicola Dollfus, and physiology in determining life history patterns Chabaud & Golvan, 1957 on the fecundity and survival in Atlantic salmon (Salmo salar). Canadian Journal of of Pomatoschistus microps (Kroyer, 1838) (Teleostei: Fisheries and Aquatic Sciences 55 (Suppl. 1), 93–103. Gobiidae). Parasitology 119, 61–7. Milinski, M. and Bakker, Th.C.M. (1990) Female stickle- Partridge, L. (1988) The rare-male effect: what is its evo- backs use male coloration in mate choice and hence lutionary significance? Philosophical Transactions of avoid parasitized males. Nature 344, 330–3. the Royal Society Series B 319, 525–39. Miller, R.M. (1957) Have the genetic patterns of fishes Pauly, D. (1980) On the interrelationships between nat- been altered by introductions or by selective fishing? ural mortality, growth parameters, and mean environ- Journal of the Fisheries Research Board of Canada 14, mental temperature in 175 fish stocks. Journal du 797–806. Conseil, Conseil International pour l’Exploration de Murphy, G.I. (1968) Pattern in life history and the envi- la Mer 39, 175–92. ronment. American Naturalist 102, 391–403. Pepin, P. (1991) The effect of temperature and size on Myers, R.A. (1984) Demographic consequences of preco- development, mortality, and survival rates of the cious maturation of Atlantic salmon (Salmo salar). pelagic early life history stages of marine fish. Canadian Journal of Fisheries and Aquatic Sciences Canadian Journal of Fisheries and Aquatic Sciences 41, 1349–53. 48, 503–18. Myers, R.A. (1986) Game theory and the evolution of Pepin, P. (1993) Application of empirical size-dependent Atlantic Salmon (Salmo salar) age at maturation. In: models of larval fish vital rates to the study of pro- D. Meerburg (ed.) Salmonid Age at Maturity. Ottawa: duction: accuracy and association with adult stock Department of Fisheries and Oceans, Canadian Special dynamics in a comparison among species. Canadian Publication of Fisheries and Aquatic Sciences 89, pp. Journal of Fisheries and Aquatic Sciences 50, 53–9. 53–61. Pepin, P., Shears, T.H. and Lafontaine, D.E. (1992) Signifi- Myers, R.A. and Hutchings, J.A. (1986) Selection against cance of body size to the interaction between a larval parr maturation in Atlantic salmon. Aquaculture 53, fish (Mallotus villosus) and a vertebrate predator (Gas- 313–20. terosteus aculeatus). Marine Ecology Progress Series Myers, R.A. and Hutchings, J.A. (1987a) A spurious corre- 81, 1–12. lation in an interpopulation comparison of Atlantic Pettifor, R.A., Perrins, C.M. and McCleery, R.H. (1988) salmon life histories. Ecology 68, 1839–43. Individual optimization of clutch size in great tits. Myers, R.A. and Hutchings, J.A. (1987b) Mating of Nature 336, 160–2. anadromous Atlantic salmon, Salmo salar L., with Policansky, D. (1993) Fishing as a cause of evolution mature male parr. Journal of Fish Biology 31, 143–6. in fishes. In: T.K. Stokes, J.M. McGlade and R. Law Myers, R.A., Hutchings, J.A. and Gibson, R.J. (1986) (eds) The Exploitation of Evolving Resources. Berlin: Variation in male parr maturation within and Springer-Verlag, pp. 2–18. among populations of Atlantic salmon, Salmo salar. Power, G. (1969) The Salmon of Ungava Bay. Montréal: Canadian Journal of Fisheries and Aquatic Sciences McGill University, Arctic Institute of North America 43, 1242–8. Technical Paper 22. Naevdal, G., Holm, M., Moller, D. and Osthus, O.D. Pressley, P.H. (1981) Parental effort and the evolution of (1976) Variation in growth rate and age at sexual matu- nest-guarding tactics in the threespine stickleback, rity in Atlantic salmon. International Council for the Gasterosteus aculeatus L. Evolution 35, 282–95. Exploration of the Sea C.M.1976/E:40. Copenhagen: Quinn, T.P., Hendry, A.P. and Wetzel, L.A. (1995) The in- ICES. fluence of life history trade-offs and the size of incuba- Näslund, I., Degerman, E. and Nordwall, F. (1998) Brown tion gravels on egg size variation in sockeye salmon trout (Salmo trutta) habitat use and life history in (Oncorhynchus nerka). Oikos 74, 425–38. Swedish streams: possible effects of biotic interac- Reynolds, J.D., Jennings, S. and Dulvy, N.K. (2001) Life Life Histories 173

histories of fishes and population responses to ex- Schaffer, W.M. (1974a) Selection for optimal life histo- ploitation. In: J.D. Reynolds, G.M. Mace, K.H. Redford ries: the effects of age structure. Ecology 55, 291–303. and J.G. Robinson (eds) Conservation of Exploited Schaffer, W.M. (1974b) Optimal reproductive effort in Species. Cambridge: Cambridge University Press, pp. fluctuating environments. American Naturalist 108, 147–68. 783–90. Reznick, D. (1985) Costs of reproduction: an evaluation Schaffer, W.M. and Elson, P.F. (1975) The adaptive signifi- of the empirical evidence. Oikos 44, 257–67. cance of variations in life history among local popula- Reznick, D.N. and Yang, T. (1993) The influence of fluc- tions of Atlantic salmon in North America. Ecology tuating resources on life history: patterns of allocation 56, 577–90. and plasticity in female guppies. Ecology 74, 1011–29. Schultz, E.T. and Conover, D.O. (1999) The allometry of Reznick, D.N., Bryga, H. and Endler, J.A. (1990) energy reserve depletion: test of a mechanism for size- Experimentally-induced life history evolution in a dependent winter mortality. Oecologia 119, 474–83. natural population. Nature 346, 357–9. Seger, J. and Brockmann, H.J. (1987) What is bet-hedging? Ricker, W.E. (1981) Changes in the average size and aver- Oxford Surveys in Evolutionary Biology 4, 182–211. age age of Pacific salmon. Canadian Journal of Fish- Simpson, B.R.C. (1979) The phenology of annual killi- eries and Aquatic Sciences 38, 1636–56. fishes. Symposium of the Zoological Society of Rijnsdorp, A.D. (1993) Fisheries as a large-scale experi- London 44, 243–61. ment on life-history evolution: disentangling pheno- Sinderman, C.J. (1987) Effects of parasites on fish popula- typic and genetic effects in changes in maturation tions: practical considerations. International Journal and reproduction of North Sea plaice, Pleuronectes of Parasitology 17, 371–82. platessa L. Oecologia 96, 391–401. Smith, C.C. and Fretwell, S.D. (1974) The optimal bal- Rijnsdorp, A.D. and Ibelings, B. (1989) Sexual dimor- ance between size and number of offspring. American phism in the energetics of reproduction and growth Naturalist 108, 499–506. of North Sea plaice, Pleuronectes platessa. Journal of Svärdson, G. (1949) Natural selection and egg number in Fish Biology 35, 401–15. fish. Report of the Institute of Freshwater Research, Rijnsdorp, A.D. and Vingerhoed, B. (1994) The ecological Drottningholm 29, 115–22. significance of geographical and seasonal differences in Taborsky, M. (1994) Sneakers, satellites, and helpers: par- egg size in sole, Solea solea (L). Netherlands Journal of asitic and cooperative behavior in fish reproduction. Sea Research 32, 255–70. Advances in the Study of Behavior 23, 1–100. Rochet, M.-J. (1998) Short-term effects of fishing on life Thomaz, D., Beall, E. and Burke, T. (1997) Alternative re- history traits of fishes. ICES Journal of Marine Science productive tactics in Atlantic salmon: factors affecting 55, 371–91. mature parr success. Proceedings of the Royal Society Roff, D.A. (1984) The evolution of life history parameters of London Series B 264, 219–26. in teleosts. Canadian Journal of Fisheries and Aquatic Thorpe, J.E. (1986) Age at first maturity in Atlantic Sciences 41, 989–1000. salmon, Salmo salar: freshwater period influences Roff, D.A. (1991) The evolution of life-history variation and conflicts with smolting. In: D. Meerburg (ed.) in fishes, with particular reference to flatfishes. Salmonid Age at Maturity. Ottawa: Department of Netherlands Journal of Sea Research 27, 197–207. Fisheries and Oceans, Canadian Special Publication of Roff, D.A. (1992) The Evolution of Life Histories. New Fisheries and Aquatic Sciences 89, pp. 7–14. York: Chapman & Hall. Thorpe, J.E., Morgan, R.I.G., Talbot, C. and Miles, M.S. Rowell, C.A. (1993) The effects of fishing on the timing of (1983) Inheritance of developmental rates in Atlantic maturity in North Sea cod (Gadus morhua L.). In: T.K. salmon, Salmo salar L. Aquaculture 33, 119–28. Stokes, J.M. McGlade and R. Law (eds) The Exploita- Thorpe, J.E., Mangel, M., Metcalfe, N.B. and tion of Evolving Resources. Berlin: Springer-Verlag, Huntingford, F.A. (1998) Modeling the proximate basis pp. 44–61. of salmonid life-history variation, with application to Sargent, R.C., Taylor, P.D. and Gross, M.R. (1987) Atlantic salmon, Salmo salar L. Evolutionary Ecology Parental care and the evolution of egg size in fishes. 12, 581–99. American Naturalist 129, 32–46. Trippel, E.A. (1998) Egg size and viability and seasonal Saunders, M.W. and McFarlane, G.A. (1993) Age offspring production of young Atlantic cod. Transac- and length at maturity of the female spiny dogfish, tions of the American Fisheries Society 127, 339–59. Squalus acanthias, in the Strait of Georgia, British Trippel, E.A., Buzeta, M.-L. and Hunt, J.J. (1995) Evalua- Columbia, Canada. Environmental Biology of Fishes tion of the cost of reproduction in Georges Bank 38, 49–58. Atlantic cod (Gadus morhua): utility of otolith 174 Chapter 7

back-calculation. In: D.H. Scor, J.M. Dean and S.E. Warner, R.R. (1991) The use of phenotypic plasticity in Campana (eds) Recent Developments in Fish Otolith coral reef fishes as tests of theory in evolutionary ecol- Research. Columbia, SC: University of South Carolina ogy. In: P.F. Sale (ed.) The Ecology of Fishes on Coral Press, pp. 599–610. Reefs. San Diego, CA: Academic Press, pp. 387–98. Trippel, E.A., Morgan, M.J., Fréchet, A. et al. (1997) Wiegmann, D.D. and Baylis, J.R. (1995) Male body Changes in age and length at sexual maturity of north- size and paternal behaviour in smallmouth bass, west Atlantic cod, haddock and pollock stocks, Micropterus dolomieui (Pisces: Centrarchidae). Ani- 1972–1995. Ottawa: Department of Fisheries and mal Behaviour 50, 1543–55. Oceans, Canadian Technical Report of Fisheries and Williams, G.C. (1966) Natural selection, the cost of Aquatic Sciences 2157. reproduction, and a refinement of Lack’s principle. Tyler, C.R. and Sumpter, J.P. (1996) Oocyte growth and American Naturalist 100, 687–90. development in teleosts. Reviews in Fish Biology and Winkler, D.W. and Wallin, K. (1987) Offspring size Fisheries 6, 287–318. and number: a life history model linking effort per Uglem, I., Rosenqvist, G. and Wasslavik, H.S. (2000) offspring and total effort. American Naturalist 129, Phenotypic variation between dimorphic males in 708–20. corkwing wrasse. Journal of Fish Biology 57, 1–14. Winterbottom, R. and Emery, A.R. (1981) A new genus van den Berghe, E.P. (1990) Variable parental care in a and two new species of gobiid fishes (Perciformes) from labrid fish: how care might evolve. Ethology 84, the Chagos Archipelago, central Indian Ocean. Envi- 319–33. ronmental Biology of Fishes 6, 139–49. van den Berghe, E.P. and Gross, M.R. (1989) Natural Wootton, R.J. (1994) Life-histories as sampling devices: selection resulting from female breeding competition optimum egg size in pelagic fishes. Journal of Fish Biol- in a Pacific salmon (coho: Oncorhynchus kisutch). ogy 45, 1067–77. Evolution 43, 125–40. Wootton, R.J. (1998) Ecology of Teleost Fishes. Vøllestad, L.A. and L’Abée-Lund, J.H. (1994) Evolution of Dordrecht: Kluwer Academic Publishers. the life history of Arctic charr, Salvelinus alpinus. Evo- Zastrow, C.E., Houde, E.D. and Saunders, E.H. (1989) lutionary Ecology 8, 315–27. Quality of striped bass (Morone saxatilis) eggs in rela- Walkey, M. and Meakins, R.H. (1970) An attempt to tion to river source and female weight. Rapports et balance the energy budget of a host–parasite system. Procès-verbaux des Réunions, Conseil International Journal of Fish Biology 2, 361–72. pour l’Exploration de la Mer 191, 34–42. Walsh, S.J. and Morgan, M.J. (1999) Variation in matura- Zimmerer, E.J. and Kallman, K.D. (1989) Genetic basis tion of yellowtail flounder (Pleuronectes ferruginea) for alternative reproductive tactics in the pygmy on the Grand Bank. Journal of Northwest Atlantic swordtail Xiphophorus nigrensis. Evolution 43, Fishery Science 25, 47–59. 1298–307. 8 Migration

JULIAN METCALFE, GEOFF ARNOLD AND ROBERT McDOWALL

8.1 INTRODUCTION migration (e.g. Harden Jones 1968; Baker 1978; McKeown 1984; McDowall 1988) or to individual Migration is a common feature of the life histories species (e.g. Pickett and Pawson 1994). Our aim is of many of the world’s economically important to review the occurrence of migration in the life species of fish. For many, these movements are histories of commercial fish species and the eco- extensive, and the regular seasonal changes in logical principles that drive migratory behaviour. population distribution that result often have We describe a number of different migratory strate- significant implications for commercial fisheries. gies and attempt to explain how understanding mi- Consequently, understanding migration is impor- gration is of value to fisheries management and tant to the cost-effective and sustainable exploita- conservation. Many species of fish live in environ- tion of fish stocks. ments that are hostile, at least to human observers, ‘Understanding’ in this context may simply in- and their migrations are therefore particularly dif- volve an appreciation of the seasonal arrival and ficult to study, so finally we review the field tech- disappearance of a stock, allowing fishermen to niques used to study fish migration and consider target their activities effectively. Alternatively, it how these are likely to develop and improve our may involve a more detailed knowledge that al- understanding in the future. lows the development of realistic models of popu- lation dynamics that can be used to assess the 8.1.1 Definitions likely effects of different management strategies. However, fish migration is a very broad subject. At ‘Migration’ is a word in common parlance and one end it involves the behavioural ecology of most of us, lay person or scientist, have a sense of whole populations, at the other it focuses on the what it means. Nonetheless, it is difficult to find a neurobiology of sensory systems that form the generally agreed definition and there are almost as basis of the orientation and navigation abilities of many interpretations as there are books on the sub- individuals. ject. Those who work on different animal groups While the different aspects of migration fasci- tend to use the term in different ways. Ornitholo- nate the biologist, not all are directly relevant in a gists usually regard migration as some form of fisheries context, and some are beyond the limited regular, long-range movement between wintering scope of this chapter. Further, it is not our aim to and breeding areas, and expect an element of give exhaustive accounts of the life-history strate- seasonal ‘to and fro-ness’ of both individuals gies, or migration circuits, of all economically and populations. This is quite different from the important migratory fish. For this the reader is entomologist, who has little expectation that directed to other, more extensive texts relating to populations, let alone individuals, of migratory 176 Chapter 8 aphids or locusts will return to the place from impact on the activities of fishermen, then such which they departed. migrations may not be very important in a fish- While there is endless debate over the details of eries context. Alternatively, migrations that cover various definitions, there is general agreement large distances will be a crucial factor determining that ‘migration’ involves the movement of indi- the seasonal changes in the distribution of fish viduals and populations from one, usually well- stocks and therefore of the fisheries which exploit defined, area or habitat to another. Furthermore, them. migration is usually understood to imply that there is some cyclical element to the movement, 8.1.3 Terminology on either an annual or life-cycle scale, which dis- tinguishes it from ‘dispersal’. Though possibly In addition to defining migration, there is a need to rather broad, such a definition is quite adequate understand some of the terms used to describe fish for the purposes of this chapter, and there is little migrations, particularly with respect to the move- value in debating what is, and what is not, ‘true’ ments of individuals within and between freshwa- migration. For fuller discourse on this subject, the ter and marine habitats. Fortunately, there is much reader can refer to a variety of books (e.g. Harden less debate about these definitions. The terms pro- Jones 1968; Baker 1978; McKeown 1984; Dingle posed by Myers (1949) and used widely since (e.g. 1996). Harden Jones 1968; McDowall 1988) are used here It should be noted that many species of fish also (Fig. 8.1). make regular vertical movements, usually linked Oceanodromous: ‘Truly migratory fishes whose to the day–night cycle, that are referred to as ‘mi- migrations occur wholly within the sea.’ gration’. Such movements certainly have implica- Potamodromous: ‘Truly migratory fishes whose tions for fisheries, particularly with respect to the migrations occur wholly within fresh water.’ availability of fish to fishing or survey gear. How- Diadromous: ‘Truly migratory fishes which mi- ever in this chapter we restrict our definition to grate between the sea and fresh water.’ horizontal migrations. First, however, it is neces- When Myers (1949) first introduced the term sary to consider scales of migratory movement. ‘diadromy’, he also defined three subcategories: ‘anadromy’ and ‘catadromy’, which had been in 8.1.2 Scales of movement long usage in the fish literature (Meek 1916), and ‘amphidromy’, which was a new term. These sub- Some littoral species, like blennies (family categories refer to various forms of diadromy, in Blenniidae), make seasonal inshore and offshore which the directions of movement and the life- movements that extend no more than a few kilo- history stages that undertake them vary. Recently, metres. In temperate waters, species like herring these categories of diadromy have been expanded (Clupea harengus), mackerel (Scomber scombrus), and refined (McDowall 1998a). cod (Gadus morhua) and plaice (Pleuronectes 1 Anadromy refers to diadromous fishes in which platessa) make more extensive movements over most feeding and growth take place at sea, prior to several hundreds of kilometres. Finally, some the migration of fully grown adult fish into fresh species migrate over distances of several thou- water to reproduce; there is either no subsequent sands of kilometres. Examples include diadro- feeding in fresh water or any feeding is accompa- mous species like Atlantic salmon (genus Salmo), nied by little somatic growth; the principal feeding Pacific salmon (genus Oncorhynchus) and eels and growing biome (the sea) differs from the repro- (Anguilla spp.), which move between fresh water ductive biome (fresh water). Examples are salmon, and the open sea, and the various species of tuna, shad and sea-lamprey. billfishes and large sharks that make extensive 2 Catadromy refers to diadromous fishes in which transoceanic migrations. Clearly, if the distance most feeding and growth occurs in fresh water moved is small (say a few kilometres) and does not prior to the migration of fully grown adult fish to Migration 177

Fresh Anadromy water Reproduction Reproduction Early feeding and growth in fresh water

Larval/juvenile migration to sea Adult return migration to fresh water

Sea Most feeding and growth in the sea

Catadromy Fresh water Most feeding and growth in fresh water

Adult return migration to the sea Juvenile migration to fresh water Reproduction

Sea Early feeding and growth at sea Reproduction

Amphidromy Reproduction Reproduction Fresh Most feeding and growth in fresh water water

Larval migration to sea

Juvenile migration back to fresh water

Sea Early feeding and growth at sea

Fig. 8.1 Different forms of diadromy. 178 Chapter 8 sea to reproduce; there is either no subsequent cal variables, such as depth, water currents, topog- feeding at sea or any feeding is accompanied by raphy, sediment type, temperature, salinity, oxy- little somatic growth; the principal feeding and gen, illumination and biotic variables, such as food growing biome (fresh water) differs from the availability and predator abundance. In freshwater reproductive biome (the sea). Examples are Ameri- lakes and rivers, there are similar, but probably can and European eels. more apparent, differences in habitat type. 3 Amphidromy refers to diadromous fishes in Such features can vary widely spatially and which there is a migration of larval fish to sea temporally, particularly between seasons. With soon after hatching, followed by early feeding and such diversity, it is very unlikely that any one habi- growth at sea, and then a migration of small post- tat will be equally suitable for all life-cycle stages, larval or juvenile fish from the sea back into fresh and many fish species do indeed show some onto- water; there is further prolonged feeding in fresh genetic and/or seasonal changes in habitat use. A water, during which most somatic growth from ju- habitat with plenty of small prey and few large venile to adult stages occurs, as well as maturation predators may be ideal for small juvenile fish, but and reproduction; there is usually no migration is less likely to be suitable for larger adults of the related to reproduction, and the principal feeding same species that need larger prey and may be less biome is the same as the reproductive biome (fresh vulnerable to predation. To make optimum use of water). An example is the ayu (Plecoglossus al- such environmental diversity, many fish species tivelis) from Japan. have evolved migratory life histories and move be- tween different areas, each of which is best for the activities of feeding, growing and spawning. As a 8.2 EXPLOITATION result, the size of a population is not limited to the AND ECOLOGY number of individuals that can be supported by a 8.2.1 Migration among single habitat. Improved survival and reproductive success is the ultimate selective advantage of mi- commercial species gratory behaviour. Worldwide, there are approximately 25000 species of fish (Eschmeyer 1998), of which probably only 8.2.3 Migration and exploitation 200–300 make extensive migrations (Harden Jones 1980). Yet, while migration is not a common fea- Migrating fish tend to move along known path- ture for most fish species, it nonetheless seems to ways at known times, and so are more concen- be widespread in the life histories of most econo- trated both spatially and temporally and hence mically important species. For example, in 1997, easier to catch. This latter aspect applies particu- which is the most recent year for which FAO data larly to diadromous fish, which are targeted by are available, the average world catch of finfish, ex- fisheries as they migrate through narrow bottle- cluding shellfish, was about 79.5 million tonnes necks at the mouths of rivers. When it comes to the from over 740 species. Of this total, the top 40% fisheries exploitation of the very small amphidro- (32.1 million tonnes) comprises only 15 species, all mous fishes, which tend to return to fresh water of which are considered to be migratory (Table 8.1). from the sea at lengths as small as 20–25mm, it is So why is migration such a common feature of eco- only their spatiotemporal concentration at the nomically important species? time of migration that makes fishing economic. Thus there are significant fisheries for the tiny mi- 8.2.2 Why do fish migrate? gratory larvae of diverse goby species in many dif- ferent places, such as the Philippines, Tahiti, the One bit of sea often appears very much like any Caribbean, Sri Lanka and Réunion. There are also other, but below the surface there is tremendous fisheries for galaxiids in Australia, New Zealand diversity. Marine habitats are a complex of physi- and Chile (reviewed in McDowall 1988). The same Migration 179

Table 8.1 FAO finish catch data for 1997 for the top 15 most important species with evidence for migratory behaviour.

Species Catch Catch as proportion Cumulative Evidence for migration (million tonnes) of total worldwide catch (%) catch of all finfish (%)

Anchoveta (Peruvian anchovy), Engraulis 7685098 9.66 9.66 Ben-Tuvia (1995) ringens Alaska (walleye) pollock, Theragra 4368108 5.49 15.16 Hamatsu and Yabuki (1995) chalcogramma Chilean jack mackerel, Trachurus murphyi 3597117 4.52 19.68 Serra (1991) Atlantic herring, Clupea harengus 2532525 3.18 22.86 Harden Jones (1968), McGladdery and Burt (1985), Creaser et al. (1984) Chub mackerel, Scomber japonicus 2423235 3.05 25.91 Belyaev (1984), Novikov (1986) Japanese anchovy, Engraulis japonicus 1666503 2.10 28.00 Novikov (1986), Wu (1989) Capelin, Mallotus villosus 1604555 2.02 30.02 Carscadden (1984) Skipjack tuna, Katsuwonus pelamis 1425638 1.79 31.82 Gauldie and Sharp (1996), Watanabe et al. (1995) Atlantic cod, Gadus morhua 1362174 1.71 33.53 Harden Jones (1968), Malmberg (1980), Rose (1993) Largehead hairtail, Trichiurus lepturus 1201491 1.51 35.04 Lin (1987) Yellowfin tuna, Thunnus albacares 1128377 1.42 36.36 Suzuki et al. (1978), Bard et al. (1987) European pilchard (sardine), Sardina 1031467 1.30 37.75 Tosello Bancal (1989) pilchardus South American pilchard, Sardinops sagax 721711 0.91 38.66 Novikov (1986), Wu (1989) European sprat, Sprattus sprattus 701202 0.88 39.54 Heincke (1992), Hoziosky et al. (1989) Blue whiting, Micromesistius poutassou 697597 0.88 40.42 Carre and Geistdoerfer (1981) Total worldwide catch of the top 15 finfish 32146798 species Total worldwide catch of all finfish species 79530900

principle applies to the exploitation of the juvenile of a selection of migratory species to illustrate dif- life stages of catadromous fishes as they enter fresh ferent factors driving their migrations. water. Glass-eels can be harvested most efficiently and economically when they are concentrated, usually at the mouths of rivers or in estuaries. An 8.3.1 Oceanodromous fish example is the elver fishery in the River Severn in Fish movements and water currents England (Templeton and Churchward 1990). Although the adults of many open-ocean species are large or powerful swimmers, individuals of 8.3 FISH MIGRATIONS most species are, either as eggs and/or free- swimming larvae, at the mercy of prevailing water The range of migratory life-history strategies is ex- currents, which therefore play an important part tensive. In this section we describe the migrations in the life cycle of most species. The idea that the 180 Chapter 8

distribution and migrations of a fish stock could be 8.3) are very much stronger (50–100cms-1 or more) linked by a particular water-current system gave and extend to depths of 1000m. In many places, rise to the concept of ‘hydrographic containment’ there are substantial countercurrents below the (Cushing 1990) and it appears that stocks of a surface circulation (Harden Jones 1968; Arnold number of important commercial species such as 1981). plaice, herring and some species of tuna could be There is compelling evidence to indicate that contained in this way. ocean currents play a major role in the movements In the open sea, as distinct from estuaries, two of larvae of several species of eel from their spawn- distinct current regimes can be recognized. Ocean ing areas in the open ocean to the freshwater habi- currents dominate on the high seas, whereas tidal tats where the juveniles feed, grow and develop currents are more important on the continental into adults. Schmidt (1922, 1923) was the first shelves. In the main ocean basins the effects of the to publish an account of the migrations of the tides are slight and the surface circulation is driven European eel Anguilla anguilla from its spawning largely by the prevailing planetary wind system grounds in the Sargasso Sea. Subsequent analysis (Fig. 8.2a). This generates subtropical gyrals that of the hydrography of the North Atlantic has circulate clockwise in the Northern Hemisphere shown that the eel’s leptocephalus larva is trans- and anticlockwise in the Southern Hemisphere ported by the Gulf Stream and North Atlantic (Fig. 8.2b). These gyrals are bounded on their Drift (Fig. 8.3) to the European continental shelf northern and southern edges by zonal currents, over a period of nearly 3 years (Harden Jones 1968). which flow broadly east or west. While open-ocean More recently, the location of the spawning area of currents are generally rather slow (3–7cms-1) and the Japanese eel Anguilla japonica has been identi- extend to depths of 100–200m, western boundary fied in the northwestern Pacific Ocean in an area currents such as the Gulf Stream in the North At- between the Philippines and the Mariana Islands lantic and the Kuroshio in the North Pacific (Fig. (Tsukamoto 1992). This discovery suggests a simi-

(a)90 N (b) 90 N 60 60 Subpolar Gyre Westerlies 30 30 Subtropical Gyre Land Land LandNE Land NE Trade Winds EC 0 0 Doldrums SE

SE Trade Winds Subtropical Gyre 30 30

West Wind Drift Westerlies 60 60

90 S 90 S

Fig. 8.2 Generalized atmospheric (a) and oceanic (b) circulations. NE, North Equatorial Current; EC, Equatorial Counter Current; SE, South Equatorial Current. (Source: redrawn from Gross 1972. Reproduced by permission of Prentice-Hall.) Migration 181

140° 160° 180° 160° 140° 120° 100° 80° 60° 40° 20° 0° 20° 40° 60° 80° 100° 120° 80°

22

c ti n a 60° tl . A N

West wind m 40° dri trea o ft lf s oshi Gu Kur C A 20°

0°

20° D E B 40°

60°

80°

Fig. 8.3 Simplified chart of the surface ocean currents of the world oceans. The main subtropical gyrals are A, North Pacific; B, South Pacific; C, North Atlantic; D, South Atlantic; E, South Indian. Selected main ocean currents are also indicated. Important areas of upwelling in the eastern boundary currents are shown as stippled areas. (Source: redrawn from Bramwell 1977.)

lar role for the North Equatorial and Kuroshio (Fig. 8.4) in which the water flows in opposite di- currents (Fig. 8.3) as a means of transporting rections for successive periods of about 6.25h. leptocephali of this species to the coastal waters During neap tides, peak tidal stream speeds can of eastern Asia, where the adults originate. range from as little as a few centimetres per second In shallow seas on continental shelves, water in the central North Sea to about 1ms-1 in the movements are dominated by tidal currents. In Dover Strait. During spring tides, peak tidal shelf seas, the constraining effects of ocean basin stream speeds increase to between 0.1 and 2.0ms-1 geometry and the influence of the Coriolis force re- respectively. sult in development of amphidromic systems in Whether in the open ocean or in shelf seas, most which tidal waves rotate around amphidromic teleost fish spawn in midwater and their pelagic points at which there is no rise and fall in water eggs and larvae usually drift downstream with the level. On the European continental shelf, for ex- residual current. At some later stage in their life ample, the tidal wave rotates approximately once history, the adults must make some compensatory every 12.5h, giving rise to a system of tidal streams movement if the population is to remain within 182 Chapter 8

8° 6° 4° 2° 0° 2° 4° 6° 8° 10° 60°

58°

56°

54°

52°

Velocity key (cm s–1)

50° >100 75 – 100 25 – 75 Fig. 8.4 A tidal stream chart for <25 the North Sea and adjacent areas. (Source: from Harden Jones et al. 48° 1978.)

the same hydrographic system (Harden Jones plaice (Pleuronectes platessa) that, in some areas 1968). While such movements may involve adults of the North Sea, takes advantage of the tidal cur- having to swim actively against the current, this rents to minimize the metabolic cost of migration. may not necessarily be required. In the open ocean, fish may exploit deepwater countercurrents (e.g. An open-ocean species: the bluefin tuna bluefin tuna, see below); on continental shelves fish may make selective use of the oscillating tidal Tuna and billfishes support major commercial and streams (e.g. plaice, see below). sport fisheries in temperate and tropical oceans Marine fish make up over 90% of the total throughout the world and are one of the classic world commercial finfish catch. Here we describe groups of migratory fish of the open ocean. How- the migrations of three very different species. The ever, their migration routes, areas of abundance first is a highly migratory open-ocean species, and availability can vary significantly from year to the Atlantic bluefin tuna (Thunnus thynnus). The year, and our lack of understanding of their migra- second is the European sea bass (Dicentrarchus tions is a significant source of problems in the labrax), a shelf-sea species whose seasonal migra- management of tuna fisheries. tions in UK waters appear to be a thermoregulatory North Atlantic bluefin tuna are assessed and mechanism that allows the females successfully managed on the basis that there are two distinct to develop their ovaries. The third species is the stocks: one in the west and one in the east. In the Migration 183 east, the stock is thought to have its major spawn- experiments in the Gulf of Maine with giant ing area in the Mediterranean Sea, with highest bluefin tuna tagged and released with pop-up concentrations around the Balearic Islands, satellite tags show that all 12 fish that were suc- Tyrrhenian Sea and the central Mediterranean. Fe- cessfully located during the known May–July males in the east mature earlier (4–5 years) and spawning period were in a region of the mid- achieve a smaller maximum size than their con- Atlantic bounded by Bermuda and the Azores specifics in the west. Spawning occurs in May and (Lutcavage et al. 1999). While it is not known if June. Having spawned, some fish begin a migration these fish were actually spawning, the results sug- circuit entirely within the Mediterranean but the gest that the migratory behaviour is complex and majority migrate out through the Gibraltar Strait, argue for reconsideration of existing assumptions often in shoals of up to 10000 fish. Subsequently, about North Atlantic bluefin tuna migration pat- these fish migrate north, possibly aided by the terns, spawning areas and stock structure. Even North Atlantic Drift (Fig. 8.3), the northeastern minor mixing could, in principle, have a marked limb of the North Atlantic gyre (Baker 1978). Until effect on stock assessment owing to the difference the 1920s this northerly movement seemed to ex- in population size between the two stocks. While tend no further than the English Channel. How- it is possible to model the effects of different levels ever, it now appears that the migration circuit of mixing resulting from different migratory be- extends as far north as Norway, with some individ- haviours, there are insufficient data to indicate uals travelling as far as 5000km in a few months. which model is correct. Such models are therefore Subsequently, between November and February, inadequate to provide reliable predictions about the tuna return south in deeper water and this leg how the condition of one stock may affect that of of the migration may be aided some of the time by the other (ICCAT 1999). deepwater countercurrents. In the spring, bluefin tuna return to the surface off the west coast of A shelf-sea species: the European sea bass Spain and Portugal, where they gather and exploit the small gyral currents of the area for feeding be- The European bass is found from the Mediter- fore re-entering the Mediterranean in the current ranean Sea, along the Atlantic coast north from that flows continuously eastward in the middle of Morocco to Ireland in the northwest and to Nor- the Gibraltar Strait (Rey 1983). way in the northeast. The species is exploited by The western stock is thought to have its major commercial fishermen and sports fisheries, with spawning area in the Gulf of Mexico and in the most fish caught in the southern part of its range Florida Straits, where spawning occurs from mid- (Pickett and Pawson 1994). April into June. Females, which are thought to Spawning occurs between February and June, spawn first at an age of about 8 years, can live for 20 and temperature appears to play an important part years or more, grow up to 3m in length and exceed in its timing and location. Bass eggs are rarely 600kg in weight. Depending on her size, a single fe- found in water below about 8.5–9°C (Thompson male can spawn between 1 million and 30 million and Harrop 1987). Bass spawn in midwater and eggs in a single season. Juveniles are thought to their eggs take 4–9 days to hatch, depending on occur in the summer over the continental shelf, temperature (Jennings and Pawson 1991). Over the primarily in the area from 34°N to 41°W and off- following 2–3 months the growing larvae drift in- shore of that area in the winter. Although we know shore on the prevailing current into sheltered very little about movements between the western creeks and estuaries where they feed and grow over and eastern stocks of North Atlantic bluefin tuna, the next 4–5 years. present assessment models assume constant pro- Conventional tagging studies show that juve- portional annual migration rates both west to east nile bass probably stray no further than 80km from (about 4%) and east to west (1–6%) (Punt and their release site (Pawson et al. 1987), but as they Butterworth 1994). However, results from recent approach maturity there is a marked increase in 184 Chapter 8

the scale of their seasonal movements. Broadly Channel. Some bass have been shown to move speaking, the pattern is for fish at the northern 800km from their summer feeding area around the limits of their geographical range to move south south coast of the UK to winter in the Bay of Biscay. and/or west in autumn, with a return migration The seasonal movements of adult bass appear in spring (Fig. 8.5). Fish from summer feeding to be strongly linked to environmental tempera- grounds on the west coast of the UK move south to ture and gonad development (Pawson and Pickett the Celtic Sea, while fish living in the southern 1996). Adolescent bass, classed as those that are North Sea and eastern English Channel move maturing but have yet to spawn for the first time, south and/or west to winter in pre-spawning areas will overwinter in water that falls below 9°C, but in the western English Channel. females do not develop fully mature gonads. Suc- Conventional tag returns of adults caught on cessful gonad maturation appears to depend on fish summer feeding grounds indicate that individuals remaining in water above 10°C through the winter usually return to specific feeding areas. However, and, to achieve this, fish in the more northerly part the distances moved appear to depend on where of their range need to migrate south and/or west in fish spent the summer (Pawson et al. 1987). Those autumn and winter (Pawson et al. 2000). living around Cornwall and South Devon may Such distinct seasonal movements have a pro- move less than 100km, while others living around nounced effect on the temporal and geographic Cumbria or in the southern North Sea may move distribution of the bass fishery. Bass above 36cm, distances of 400–500km to the western English which is the legal minimum landing size, are

8° 6° 4° 2° 0° 2° 8° 6° 4° 2° 0° 2° (a) (b)

54° 54° Sept June

Nov Oct Oct May ° ° 52 Sept 52 May June Nov Feb Nov Oct May June Apr 50° Nov 50° Feb Dec

48° 48°

Fig. 8.5 Typical migration patterns of adult sea bass populations around England and Wales as shown by conventional tagging experiments (solid arrows) and postulated movements (broken arrows) in (a) autumn and (b) spring. The three different forms of shading indicate different adult populations. (Source: from Pawson et al. 1987.) Migration 185 caught throughout the year in the English been established from trawl surveys (Wimpenny Channel, whereas at the northern limits of their 1953) and conventional tagging experiments in the range in the Solway Firth the main fishing season Southern Bight of the North Sea (see review by is restricted to July and August (Pawson and Harden Jones 1968) and the English Channel Pickett 1996). In the summer, bass tend to move (Houghton and Harding 1976). In autumn, matur- into shallow warmer water and much of the fish- ing plaice leave the Leman Ground and more ing is concentrated in inshore areas. In winter, northerly feeding areas and migrate south to adult fish move offshore into deeper water, and spawning areas centred on the Hinder Ground and large catches of pre-spawning bass are taken, par- in the eastern English Channel. On the Hinder ticularly by French pair-trawlers, in the western spawning grounds, peak spawning occurs in late English Channel in late winter and spring. In January and February (Simpson 1959; Harding spring and early summer fish move north, and fish- et al. 1978), whilst peak spawning in the eastern eries in the Solent, South Wales and the Thames English Channel occurs in early January Estuary peak through May and June, and again in (Houghton and Harding 1976). Subsequently, September and October when the fish return spent plaice return north to their summer feeding south. areas in late winter and spring. The timing of these The knowledge that the seasonal movements of migrations appears to be well structured both bass are tightly linked to environmental tempera- by age and sex. Extensive midwater trawling ture makes it possible to predict how climatic experiments in the Dover Strait have shown that changes may affect distribution and stock struc- males complete their pre-spawning migration ture. For instance, if winter temperatures become earlier, and their post-spawning migration later, warmer, bass that currently spend summer in the than females (Harden Jones et al. 1979; Arnold and southern North Sea, but migrate to the western Metcalfe 1996) confirming earlier suggestions English Channel in winter, might remain in the (Hefford 1909, 1916; Simpson 1959) that males southern North Sea for the whole year. If this situa- spend longer on the spawning ground than fe- tion persisted, these fish would experience a differ- males. These experiments also suggest that imma- ent pattern of exploitation than previously and ture females follow the mature fish in a ‘dummy thus constitute a separate stock for management run’ migration that may help them learn the loca- purposes. Meanwhile, the stock of bass that tion of the spawning grounds (Arnold and Metcalfe continues to overwinter in the western English 1996). Channel would no longer receive an influx of Recently, a range of sophisticated telemetric recruiting adolescent or adult fish from the south- techniques has revealed details of the mechanism ern North Sea. Bass stocks in the English Channel by which these fish migrate. Sonar tracking stud- as a whole might thus become depleted by the ies of plaice equipped with transponding acoustic existing level of fishing effort. tags (Greer Walker et al. 1978; Metcalfe et al. 1992) and experiments with electronic data storage tags (Metcalfe and Arnold 1997, 1998) have established A shelf-sea species: the North Sea plaice that these migrations are made by selective tidal The plaice is an important commercial species in stream transport. This behaviour is characterized the North Sea, being the fifth most valuable finfish by a circatidal pattern of vertical movement in to be landed by the UK fishing fleet in 1998 (most phase with the tidal streams (Fig. 8.6). The fish recent data). As plaice have been studied since the leave the sea bed and move into midwater at about early 1900s, we probably now understand the biol- the time of slack water (Greer Walker et al. 1978) ogy and migratory behaviour of North Sea plaice and swim down-tide (Metcalfe et al. 1990; Buckley better than those of any other commercial species and Arnold 2001) for the major part of the ensuing in European waters. north-going or south-going tide. As the tide turns The general pattern of plaice migration has again, the fish return to the sea bed where they re- 186 Chapter 8

0

20

54 Depth (m)

53 40 Lowestoft

52 NS 0123

Night Day

60 0 40 80 120 160 Elapsed time (h)

Fig. 8.6 The vertical movements (solid line) of a 44-cm maturing female plaice tracked in the southern North Sea in January 1991 in relation to the depth of the sea bed (broken line) and the direction of the tidal stream. The bars along the abscissa indicate the periods of north-going (N) and south-going (S) tidal streams and the periods of day and night respectively. Inset: the ground track of the same fish as it moved north between 25 and 31 January 1991. (Source: from Metcalfe et al. 1992.) main for the duration of the opposing tide. This are slow, such as the German Bight, plaice appear behaviour allows fish to move at speeds of up to to migrate by directed swimming close to the sea 25km day-1 between feeding and spawning bed rather than by using tidal streams. These re- grounds while reducing the cost of migration by sults support the idea that metabolic cost deter- between 20% (Metcalfe et al. 1990) and 40% mines the choice of migratory mechanism, and (Weihs 1978) compared with swimming continu- that plaice only exploit tidal streams where these ously over the same distance. Other species such are sufficiently fast to reduce the cost of migration, as cod (Arnold et al. 1994), sole (Greer Walker et al. thereby enhancing the energy available for repro- 1978), eels (McCleave and Arnold 1999) and possi- duction (Metcalfe et al. 1993). bly mackerel (Castonguay and Gilbert 1995) may Knowledge that plaice in the Southern Bight also use selective tidal stream to aid migration. use tidal streams for migration has allowed devel- While selective tidal stream transport can opment of computer simulation models that pre- clearly reduce the energetic cost of migration, it dict rates and scales of geographical movement by may also provide a reliable transport mechanism combining patterns of behaviour with tidal stream for fish that are otherwise unable to navigate be- vectors (Arnold and Cook 1984; Arnold and tween feeding and spawning areas. However, re- Holford 1995). Although early models (Arnold and cent evidence (Metcalfe et al. 1999) indicates that Cook 1984) relied on generalizations about fish be- in areas of the North Sea where tidal stream speeds haviour, the recent use of electronic data storage Migration 187 tags has provided the detailed information about Clupeidae (shads and herrings), Engraulidae how behaviour changes in time and space needed (anchovies), Ariidae (forktailed catfishes), to develop more realistic models. Such models Gasterosteidae (sticklebacks), Gadidae (cods), should help to predict seasonal movements of fish Percichthyidae (temperate basses) and Gobiidae populations in relation to fishing effort, and allow (gobies) (McDowall 1988). effective assessment of management options By far the best-known examples of anadromy such as marine protected areas and the potential are the various species of salmon (genera On- effectiveness of quota management (Metcalfe et al. corhynchus and Salmo; Groot and Margolis 1991). 1998, 1999; Sparre and Hart, Chapter 13, Volume These fish typically spawn in the gravels of head- 2). water streams, occasionally at low elevations. After the eggs hatch, the young, known as alevins, 8.3.2 Diadromous fish remain in the spawning gravels for some weeks before emerging to live and feed in the streams. Diadromous fish comprise a small (2.6% in 1997) Behaviour after emergence varies with species. In but important part of the world commercial catch pink or chum salmon (O. gorbuscha and O. keta), of finfish. Myers’ (1949) definitions of the three dif- there is a rapid downstream movement and the ferent kinds of diadromous migration given earlier alevins go almost immediately to sea. At the other appear distinct and explicit, but the distinctions extreme, parr of Atlantic salmon (S. salar) may between categories (e.g. potamodromy or ocean- spend several years feeding and growing in fresh odromy on the one hand and diadromy on the water before moving to sea as smolts. Duration of other) are not always clear. In particular, there are life at sea, where most growth takes place, varies many fish from diverse higher taxa that move from just 2 years in pink salmon to as many as 5–6 between fresh water and the sea but with no par- years in chinook salmon (O. tshawytscha) (Groot ticular pattern with regard to either season or life- and Margolis 1991). When salmon at sea approach history stage. There are also others that make maturity, they migrate towards land and enter the similar movements on a short time-scale, such as mouths of rivers. They move upriver over a period diurnally or tidally. These fish are not truly diadro- of weeks to months, during which time they do not mous and some movements scarcely qualify as feed and energy is transferred from the body to the migration at all. Such species are probably best maturing reproductive organs. In most species of described as facultative wanderers. Although Pacific salmon, spawning is followed inevitably by there is a continuum from facultative wanderers to death; in some other species the spent adults may more strictly diadromous fishes, we can still use recover and mature again in subsequent years. diadromy and associated terms in an informative Lampreys follow a similar life cycle. Duration and heuristic way. The same is true for the terms of the juvenile stage in fresh water is typically sev- ‘carnivore’ and ‘herbivore’ or ‘benthic’ and ‘pela- eral years, as is the marine phase, during which gic’, which are not always totally explicit but are time the submature lampreys are parasitic on ma- still useful as general descriptors. rine fishes (Hardisty and Potter 1971). A distinc- tive feature of the return and upstream migration of the southern pouched lamprey (Geotria aus- Anadromous fish tralis) is that sexual maturity is not attained until Anadromous fishes include lampreys of the about 18 months after the adult leaves the sea, and families Petromyzontidae, Geotriidae and there is no feeding in fresh water at all (Glova Mordaciidae, the Acipenseridae (sturgeons), 1995). Osmerid smelts in northern cool temperate Salmonidae (salmons, trouts, chars and white- waters (Scott and Crossman 1973) and retropinnid fishes), Osmeridae (northern smelts) and Retro- smelts in Australia and New Zealand (McDowall pinnidae (southern smelts), Aplochitonidae 1990) are also anadromous, although the duration (Tasmanian whitebait), Salangidae (icefishes), of freshwater life is brief. Adults, which are close 188 Chapter 8 to maturity when they enter fresh water, spawn tions. Upstream penetration in some rivers of and die, and the larvae move swiftly to sea after Papua New Guinea is substantial, for example up hatching. to 800km in the Fly River (Roberts 1978). An inter- esting and distinctive feature of this species is that only the males live in fresh water. The females are Catadromous fish sex-reversed males that have left fresh water and Catadromous fishes include Anguillidae (fresh- remain in estuaries and tidal reaches (Moore 1982; water eels), some Galaxiidae (galaxiids and Moore and Reynolds 1982). southern whitebaits), some Clupeidae, Engrauli- dae, Mugilidae (grey mullets), Centropomidae Amphidromous fish (snooks), Percichthyidae, Bovichtidae (southern rock cods), Scorpaenidae (scorpionfishes), Kuhli- Amphidromy occurs among members of the idae (flagfishes), Terapontidae (therapons), Lut- Galaxiidae, Aplochitonidae (South American janidae (snappers), Cottidae (sculpins), Gobiidae peladillos), Prototroctidae (southern graylings), and Pleuronectidae (flounders) (McDowall 1988). Clupeidae, Pinguipedidae (New Zealand’s torrent- The life cycles of the 15 species of anguillid eels fish), Cottidae, Eleotridae (sleepers) and Gobiidae (family Anguillidae) are very similar. Mature fish (McDowall 1988). spawn at sea, usually in tropical to subtropical Amphidromy is the least widely recognized waters and often over great oceanic depths; invari- subcategory of diadromy, but nevertheless repre- ably they die after spawning (Bertin 1956; Tesch sents a clearly distinct pattern of migration. A 1977). The eggs hatch as distinctive, leaf-shaped, typical example in the Northern Hemisphere is leptocephalus larvae that feed and grow in the sea the Japanese ayu (Okada 1960; Kawanabe 1969). for a year to two, gradually making their way back Spawning takes place in fresh water during au- to the shores of countries from which their parents tumn and the demersal eggs develop and hatch originated, probably propelled largely by oceanic there. The larvae move immediately to sea, where currents. Leptocephali metamorphose in coastal they live for several months, before returning to waters, first to become slender and typically eel- fresh water during spring at a size of 50–70mm. shaped glass eels, and then into elvers that make The fish grow and mature over the summer, before their way into rivers from the sea. As elvers, and spawning and dying the following autumn. A very subsequently as yellow eels, they gradually pene- similar life history is evident in a number of south- trate upstream, feeding and growing over a period ern galaxiids. Spawning occurs in autumn or early of many years. Age at maturity varies widely, being winter. Newly hatched larvae emigrate to sea arguably greatest in the New Zealand longfin eel where they feed and grow over the winter, before (Anguilla dieffenbachii), in which the female may returning to fresh water at a length of 40–50mm be 80 or more years old before maturing and the following spring. In the New Zealand inanga migrating to sea (Jellyman 1995). (Galaxias maculatus), which like the ayu is an an- Some mullets (family Mugilidae) are also nual species (McDowall 1990), maturation takes catadromous. They spawn at sea and the larvae place over the first summer and the adults spawn are initially marine. Small juveniles may invade and die the following autumn (McDowall 1990). freshwater rivers and remain there until maturity, However, in other amphidromous galaxiids, matu- when there is a downstream migration to the sea. ration may not take place until the second or even Details of mullet life histories after first spawning third autumn depending on sex. In these species a are poorly understood, but there may be a return substantial number of fish survive spawning and migration to fresh water of post-spawning adults may spawn repeatedly over subsequent years (McDowall 1988). (McDowall 1988, 1990). The Australian barramundi (Lates calcarifer) is Of the 250 species of diadromous fishes, about also catadromous and undertakes similar migra- half are anadromous, one-quarter are catadromous Migration 189 and one-quarter are amphidromous (McDowall (Lawson and Carey 1972) equipped with acoustic 1988, 1998a). Anadromous fishes occur most fre- tags indicate that these fish can maintain a consis- quently in cool-temperate and subpolar waters, tent heading in the open sea for hours or even days, catadromous fishes occur most frequently in moving independently of the water current. warm-temperate to tropical waters, and amphidro- The observation that fish can use external di- mous species are most common on oceanic islands rectional clues and move independently of water in tropical to cool temperate waters (McDowall currents has led to controversy as to the relative 1988). importance of environmental transport by water currents and orientated swimming in migrations, The evolution of these three forms of diadromy particularly for large and active species like may be related to optimization of life-history salmon and tuna. There are, as yet, insufficient strategies in differing geographical ranges. data available about movements of individual fish Anadromy is favoured in cooler regions where bio- in the open sea for it to be possible to resolve this logical productivity is greater in the sea than in question. However, it is likely that there is no sin- fresh water. The sea is therefore a better feeding gle answer, and the importance of each type of and growing biome because growth is more rapid, movement differs for different stocks of fish and survival is usually higher, size at maturity is larger for different life-history stages. It seems probable and fecundity is usually greater. Catadromy is that there is a trade-off between the costs of migra- favoured in warmer regions because biological pro- tion, such as increased energy expenditure and risk ductivity is greater in fresh water than in the sea, so of predation, and the benefits, which can include that the main trophic life stage occurs in fresh improved survival and reproductive success. If this water for the same reasons that anadromous fish is so, the cost–benefit trade-off for any particular spend their main trophic life stage at sea (Gross stock of fish will vary depending on the complex of et al. 1991). Amphidromy may be favoured on environmental and biological factors experienced oceanic islands because it either provides a mecha- by that particular stock. For example, in the North nism for fish to colonize the fresh waters of islands Sea plaice exploit tidal streams during their when these become habitable or favours recolo- pre- and post-spawning migrations in areas where nization of these waters following perturbations, currents are fast enough for them to be able to save such as volcanism or drought, which cause local energy (Metcalfe et al. 1990). However, more extirpation of island stream faunas (McDowall recent evidence already described suggests that, in 1995, 1998a). areas where the tidal streams are slow, plaice swim in a directed way across the sea bed instead of using the tidal streams. 8.4 MIGRATORY MECHANISMS 8.4.2 Navigation and homing 8.4.1 Orientated swimming Our initial discussion of migration implied a Although water currents clearly play a major role strong element of ‘to and fro-ness’, and therefore in migrations of some stocks of fish, particularly in a return movement during some part of the migra- the open sea (Arnold 1981), there is also evidence tory cycle. This leads to some important ques- that some species move independently of water tions. To what extent is the return movement an currents. Atlantic cod have been shown to migrate active process? To what extent does it involve across the cold (<0°C) waters of the Newfoundland homing to ‘a place formally occupied instead of Shelf by swimming along deep ‘highways’ of warm going to equally probable places’ (Gerking 1959)? (2–2.5°C) oceanic water (Rose 1993). Tracking What mechanisms are used to locate ‘home’? The studies involving blue sharks (Prionace glauca, literature on navigation and homing in fish is ex- Carey and Scharold 1990) and bluefin tuna tensive, and numerous scholarly reviews, both 190 Chapter 8 specific and general, are available (e.g. Dittman have been suggested for plaice in the North Sea and Quinn 1996). Suffice to say, there is evidence (Arnold and Metcalfe 1996), immature Arcto- for a whole range of homing and navigational abili- Norwegian cod migrating from the Barents Sea to ties in a wide variety of fish species. the Norwegian coast (Trout 1957; Woodhead 1959) At one end of the spectrum, there is clear evi- and Newfoundland cod making inshore feeding dence for very precise homing abilities in some migrations after spawning off the edge of the conti- species. For example, adults of both Pacific and nental slope (Rose 1993). If young fish learn migra- Atlantic salmon are able to return to the gravel tion routes by following older ones, loss of older bed in which they themselves were spawned fish from a stock may lead to changes in the pattern several years earlier. Such feats of homing have of migration. The migration pattern of Norwegian been shown to depend on the young fish learning spring-spawning herring (Clupea harengus) features of their ‘home’ before they leave for the changed dramatically after adult stocks were sea, and subsequently using celestial, geomagnetic depleted in the 1960s (Dragesund et al. 1980) and and olfactory clues to aid their return from the there is concern that the migrations of high seas to their natal river (Hasler 1971). Newfoundland cod may have altered, possibly In comparison, European and Japanese eels irrevocably, as a result of the severe population spawn in the open sea in locations which ensure declines due to overfishing in the 1990s (Rose that most of their progeny will eventually be 1993). carried to the continental shelves by drifting pas- From an experimental point of view, our under- sively with ocean currents. However, unlike standing of the migrations of salmon is helped by salmon, elvers arriving on the continental shelves the fact that the young fish (smolts) can be tagged have no prior experience of the freshwater habitats before they go to sea. As a result, it is possible to they are about to enter, and it seems very unlikely identify with some accuracy how many fish return that individuals can identify and ascend the to their natal river and how many arrive at spawn- same rivers that their parents left several years ing sites in other rivers. However, acquiring com- previously. parable data for fish that spawn in the open sea is For migratory fish that live entirely at sea, like much more difficult since there is not yet a tech- cod and plaice, conventional tagging studies nique for marking eggs or larvae in a way that al- provide evidence that adults return to the same lows the individual fish to be reidentified as adults. spawning grounds in successive years (de Veen Why do some fish exhibit greater precision in 1962 quoted by Harden Jones 1968 and Cushing their homing abilities than others? Presumably, 1990). Considering the vast numbers of individual the level of precision will be a function of the trade- fish involved and the comparatively small number off between enhanced reproductive success and of available spawning grounds, it is inevitable that the increased cost involved in homing to one some fish will spawn on grounds where they them- specific spawning site as opposed to another. For selves were spawned. However, there is as yet no example, the relatively small additional cost evidence, comparable with that for salmon, to sug- incurred by an adult salmon in selecting one river, gest that adults use a specific homing mechanism as opposed to another a few miles further up the to locate their natal spawning grounds. In contrast, coast, may be significantly outweighed by the there is growing evidence that in marine species higher reproductive success achieved by spawning like herring, local population persistence is en- at a site previously proved to be good. On the sured by recruiting individuals ‘learning’ migra- other hand, recent studies of the migrations of tion patterns and spawning areas from adults of Norwegian spring-spawning herring suggest that the stock to which they recruit, rather than by homing per se is not a successful strategy and that natal homing (McQuinn 1997). Such a mechanism the selection of spawning grounds depends on the would be consistent with observations of ‘dummy condition of the fish, the cost of migration and the run’ migrations in which pre-spawning recruits probability of larval survival (Slotte and Fiksen follow mature adults. Dummy run migrations 2000). The evolution of homing in some anadro- Migration 191 mous species is usually regarded as a mechanism Petersen discs used to tag plaice and other flatfish- that allows fish to return to spawn in habitats that es, while internal tags include the tiny coded wires are capable of producing large numbers of progeny. that are used for the mass marking of young However, because it limits levels of gene flow be- salmon. tween populations, homing may also be a mecha- Simple tagging can be very useful for describing nism that enables local populations of diadromous gross patterns of population movement, but the species to become adapted to local conditions method tells us very little about how fish migrate. (Dittman and Quinn 1996). Such a situation has Population movements derived from tagging been demonstrated for populations of some Pacific studies rely on commercial fishermen reporting salmon (Hendry and Quinn 1997). details of the time and location of recapture of tagged fish and the results of such studies are in- evitably an integration of both fish behaviour and 8.5 TECHNIQUES fishing activity, which confounds any analysis of population movements. Tagging data can be ad- Understanding fish migration is clearly important justed for spatial variations in fishing effort, where to rational management and conservation of fish this is known, but movements of fish into unfished stocks, but obtaining the necessary information is or unfishable areas, or changes in fish behaviour frequently difficult. Fishers can provide useful in- that alter availability or catchability, cannot easily formation about seasonal appearances and disap- be accounted for. It is only by understanding the pearances of fish, and fisheries statistics can give a movements and behaviour of individuals over general picture of seasonal changes in population short (hours and days), medium (days and weeks) distribution. However, such information rarely and long (seasons and years) time-scales that we suffices to describe the migrations of a particular can reveal the mechanisms fish use to move about. species, and much more detailed information Understanding these mechanisms allows us to be about the behaviour and movements of individu- predictive rather than simply descriptive. als and populations is needed. Since the late 1960s electronic tags that trans- Tagging, and other simple methods of marking, mit radio or acoustic signals have increasingly have been used since the mid-17th century been used to track the movements of individual (Walton and Cotton 1898) as a means of increasing free-ranging fish for limited periods. Such work our understanding of fish biology. Tagging tells has yielded substantial advances in our under- us where individual fish are at two times in standing of how some species of fish such as the their life, when caught and tagged and when plaice migrate. However, this technique is limited recaptured. If tagging and recapture are separated because, in most applications, only one fish can be by a suitable amount of time, which can be months followed at a time, each fish can only be followed or even years, conventional tagging can provide in- for a short period, often only a few days, and sea- formation on stock identity, movements, migra- going work aboard research vessels is expensive. tion rates and routes, abundance, growth and More recently, substantial advances in microelec- mortality. tronic technology have permitted the successful There are many methods for marking or tagging development of electronic ‘data storage’ or fish (see Parker et al. 1990; Jennings et al. 2001). ‘archival’ tags that are small enough to be attached Branding or fin clipping is a quick and simple way to fish. These devices record and store environ- to mark large numbers of fish, while chemical tags mental and behavioural data and, because there is such as tetracycline, an antibiotic that is deposited no need for human observers to follow the fish, specifically in calcified areas and fluoresces under now make it possible to monitor the behaviour and ultraviolet light, can easily be applied to large movements of many fish simultaneously over en- numbers of fish and remain as a permanent mark. tire migrations (Metcalfe and Arnold 1997, 1998). Alternatively, various types of tag can be attached A variety of such devices are now being used to to, or placed in, the fish. External tags include study the movements of species as diverse as plaice 192 Chapter 8

(Metcalfe et al. 1994; Metcalfe and Arnold 1997), salmon (Walker et al. 2000), tuna (Gunn 1994; 8.6 DISTRIBUTION Gunn et al. 1994; Block et al. 1998) and others. AND GENETICS Although most data storage tags currently mea- sure only simple environmental variables such as Migratory fish tend to have much wider geographi- depth from pressure, internal and external temper- cal distributions than non-migratory species, and a ature and ambient daylight, the data can nonethe- single species may exist as a number of distinct less be used to derive detailed information about populations spread over a wide area. In some in- the movements of fish. On the European continen- stances, particularly when they are separated by tal shelf, times of high and low water and tidal large distances, different populations may have range derived from pressure measurements can be sufficient integrity in the medium to long term used to locate fish whenever they remain station- that they respond independently to the effects of ary on the sea bed for a full tidal cycle or more exploitation. In such cases these populations can (Metcalfe and Arnold 1997, 1998). In the open sea, be regarded as discrete ‘stocks’, which can be man- records of ambient daylight can be used to derive aged separately. If there is little or no transfer of latitude from daylength and longitude from the individuals between these populations, they may time of local noon (Gunn et al. 1994; Hill 1994; Hill also become genetically distinct. However, the and Braun 2001; Metcalfe 2001). The development transfer of only a few individuals can often lead of further onboard sensors that can monitor to loss of genetic heterogeneity, even though ex- more complex variables, such as compass heading, ploitation of one population may have no effect on swimming speed or feeding activity, will do much the other. Genetic analysis can thus be of value to increase our understanding of migration of when it indicates that a population consists of many more species of fish (Arnold and Dewar more than one stock, but difficulties arise when it 2001; Righton et al. 2001). fails to provide evidence of stock separation (Ward, Despite such technical advances, the use of Chapter 9, this volume). data storage tags with many species remains Among marine species, for example, Atlantic limited because the prospect of the fish being cod are found in continental shelf waters through- caught and the tags returned is very low. To avoid out the North Atlantic from the eastern seaboard the need to rely on a commercial fishery and to in- of Canada and the USA to the coast of Norway and crease the probability of data recovery, a major area into the Barents Sea. However, while there are few of development has been the ‘pop-up’ tag. These obvious physical barriers to mixing, cod do not tags are attached externally and have a release form one large mixed population and a number of mechanism that causes the tag to detach from the separate stocks have been identified. Widely sepa- fish at a predetermined time and ‘pop-up’ to the sea rate populations, such as Barents Sea and Canadian surface where the data can be recovered by air- cod, have been shown to be genetically distinct, borne radio or satellite (Nelson 1978; Hunter et al. and isolation by distance is clearly a major factor 1986). Such devices are now commercially avail- maintaining stock separation (Hutchinson et al. able and are being deployed on large pelagic species 2001). Early genetic studies of North Sea cod sug- such as tuna (Block et al. 1998; Lutcavage et al. gested that this population was a single stock 1999). Although data transmission capabilities are (Jamieson and Thompson 1972). However, more currently very limited, further developments in recent studies involving microsatellite DNA this field give the prospect of much improved analysis have shown that there is significant ge- data recovery rates in the future, while further netic substructuring of the North Sea cod popula- miniaturization will allow the technology to be tion (Hutchinson et al. 2001), with three or four applied to small species. genetically distinct groups: one off Bergen, one in the Moray Firth, one off Flamborough and one in the Southern Bight. The close proximity of these Migration 193 different North Sea groups suggests that some be- inhibits the development of genetic differences be- havioural and/or environmental factors must play tween diadromous populations and thus inhibits a role in maintaining the genetic integrity of each speciation. (Hutchinson et al. 2001). Although the temporal There are several corollaries of this principle. stability of this genetic substructuring needs to One is that the phylogenetic relationships, be assessed, the results suggest that current and narrower distributions, of non-diadromous assessment of North Sea cod may need to be re- species can often be explained in terms of the re- evaluated. cent geological histories of the areas they occupy Diadromy has particular implications for the (McDowall 1996, 1998a; Mooi and Gill, Chapter 3, evolution, biogeography and ecology of fish this volume). Moreover, the existence of diadromy species, populations and communities, with gives fish species a capacity for restoration of popu- diadromous species tending to have more lations in areas perturbed by such events as glacia- widespread distributions than closely related tion (Milner 1987; Main 1989; Milner and Bailey non-migratory species. Although not all diadro- 1989; Bernatchez and Wilson 1998), volcanic erup- mous species are widely dispersed, those that are, tions (Leider 1989; McKnight and Dahm 1990; and particularly those found on both sides of the McDowall 1996), drought (McDowall 1995) and major northern temperate oceans, are diadromous. defaunation resulting from pollution. Diadromous For example, Pacific salmon of the genus On- species re-establish populations in such water- corhynchus are found on both sides of the northern ways by migration upstream from the sea once Pacific (Scott and Crossman 1973; Groot and habitats again become habitable. Another corol- Margolis 1991), whilst the Atlantic salmon is pre- lary is that the upstream distributions of diadro- sent on both sides of the Atlantic (MacCrimmon mous species may be limited by the presence of and Gots 1979). The same is true of some sturgeons natural or artificial barriers that exclude fish from of the family Acipenseridae and some smelts of the habitats located further upstream. Some diadro- family Osmeridae in northern cool temperate mous species demonstrate astonishing ability to waters (Scott and Crossman 1973). The same ap- migrate upstream. Chinook salmon in the Pacific plies in the Southern Hemisphere. Examples are northwest of North America may migrate up- the anadromous southern pouched lamprey and stream for more than 3000km (Healey 1991), the diminutive catadromous inanga, which share while Atlantic salmon are known to penetrate an extraordinarily wide distribution that involves 2500km inland (Hasler 1971). Equally astonishing all or most of western Australia, eastern Australia, are the climbing abilities of the small migratory Tasmania, Lord Howe Island, New Zealand, the juveniles of catadromous anguillid eels and Chatham Islands, southern South America and amphidromous gobies and galaxiids, which are the Falkland Islands (McDowall 1990). capable of climbing huge natural falls and the faces Within regions, distributions and the amount of dams (McDowall 1990; Fitzsimons and of within-species genetic structuring have been Nishimoto 1995). Such swimming and climbing shown to relate to diadromy. Species that have abilities are at the extreme end of a continuum. At diadromous life stages tend to demonstrate less the other end are species that do not penetrate far genetic structuring than those that do not upstream (McDowall 1993), some of which (Allibone and Wallis 1993; Bernatchez and Wilson scarcely penetrate beyond estuarine waters. These 1998; McDowall 1999). This difference is attrib- more limited movements are presumably a combi- uted to the capacity for gene flow amongst separate nation of poor swimming or climbing ability and a populations of diadromous species. Gene flow is low instinctive drive to move far upstream. inhibited in non-diadromous species, whose popu- Where regional freshwater fish faunas are domi- lations are geographically isolated in separate nated by diadromous species, there is a variation in freshwater habitats. These differences have impli- species richness along the length of river systems, cations for patterns of speciation. Low gene flow with greatest richness at the downstream end. 194 Chapter 8

Thus, as one moves up the length of a river, the pat- acoustic surveys, during which it is normally as- tern of species richness closely resembles the sumed that the fish are stationary (but see Sparre change in species richness predicted by the ‘river and Hart, Chapter 13, Volume 2). However, this is continuum concept’ (Vannote et al. 1980; Minshall rarely true, and migration can be a major source et al. 1985). However the similar patterns of of error in acoustic surveys (MacLennan and species richness have quite different causal mech- Simmonds 1991). By properly understanding the anisms (McDowall 1998b). Running parallel to de- spatial and temporal pattern of migrations, sur- clining species richness as one moves up a river, veys may be appropriately adapted or the effects of there is a decline in fish abundance, driven by the migration on the abundance estimates otherwise same causes. Moreover, the fact that fish faunas in taken into account. rivers are dominated by diadromous species means Migration is also a key factor in the identifica- that the assembly of fish communities (Drake tion of stock units in relation to exploitation. For 1990, 1992) is, in substantial measure, driven ex- instance, a particular species may be caught in one ternally by the processes that control upstream in- area at one time of the year and in another area at vasion. Consequently, the community existing at another time. Such a situation could result either any point up a river system is the product of inter- from one stock migrating between the two areas actions between local characteristics that deter- and being exploited twice, or from two separate mine which species the habitat is suited for and stocks migrating in from two other areas at differ- those factors such as distance and obstacles that ent times. Understanding stock identity is there- determine which species are able to reach such fore important in predicting the effects of changes habitats. in the levels and/or patterns of exploitation. Indices of biotic integrity (IBI) are increasingly Similarly, migration is a critical element in the being used in freshwater ecology and resource effectiveness of management by ‘closed areas’ or management. Karr (1991) defined biological in- ‘no-take zones’ (Polunin, Chapter 14, Volume 2). tegrity as a ‘balanced, integrated community of or- Clearly, while a fish remains within a closed area it ganisms having a species composition comparable is protected from fishing, but once it moves out- to that of a natural habitat of the region’. Karr side it becomes vulnerable. It is therefore neces- (1991) used attributes such as species composition, sary to understand the spatial and temporal species richness, fish abundance, trophic composi- patterns of fish movement and fishing effort if we tion, fish reproductive guilds and fish condition to are realistically to assess the conserving effect of evaluate habitat quality and anthropogenic ef- this method of management. fects. However, where fish faunas are strongly in- fluenced by invasion through diadromy, there is a natural downstream–upstream attrition of the fish 8.8 CONCLUSIONS fauna. This means that depending on how far up- stream they are and what impediments to migra- As world fisheries continue to be heavily exploit- tion are present, similar habitats may have very ed, with drastic reductions in catches or even clo- different fish faunas. Consequently, use of the IBI sures of entire fisheries (such as the Newfoundland in river systems dominated by diadromous species cod in the 1990s) being necessary to conserve is likely to fail, unless multipliers can be developed stocks, there is an increasing need for rational that compensate for the effects of distance inland management that takes more account of funda- and the degree of difficulty in reaching various mental biology. This applies not only to tradition- habitats under study (McDowall and Taylor 2000). ally exploited species like cod and tuna but also to newly developing commercial fisheries, like those for deepwater species such as orange roughy 8.7 FISHERY APPLICATIONS (Hoplostethus atlanticus) and round-nosed grenadier (Coryphaenoides rupestris). For most of Stock assessment involves routine fishing or these species we know little of their migratory be- Migration 195 haviour or of the environmental factors that affect Holford, B.H. (1994) Vertical movements of cod it. (Gadus morhua) in relation to tidal streams in the Fortunately, management agencies are becom- southern North Sea. ICES Journal of Marine Science 51, 207–32. ing increasingly aware of the need to understand Baker, R.R. (1978) The Evolutionary Ecology of Animal fish migration, not just because it is interesting but Migration. London: Hodder and Stoughton. because it is fundamental to many basic elements Bard, F.X., Kothias, J.B.A. and Holzapfel, E. (1987) that underpin fisheries management. It is hoped Migration transatlantique d’albacore (Thunnus al- that this will lead to improved future support for bacares). Collective Volume of Scientific Papers. Inter- the study of fish migration. However, the need to national Commission for the Conservation of understand migration is of little value without the Atlantic Tunas 26, 27–30. Belyaev, V.A. (1984) The range of Japanese mackerel. necessary tools, so continuing to improve method- Okeanologiya 24, 689–95. ologies will be important too. The successful ap- Ben Tuvia, A. (1995) Biological characteristics of middle plication of telemetry techniques to the study of sized pelagic fishes. In: C. Bas, J.J. Castro and J.M. fish migration should help to ensure that, as tech- Lorenzo (eds) International Symposium on Middle- nology develops, smaller, cheaper, more sophisti- sized Pelagic Fish. Barcelona: Scientia Marina, pp. cated electronic tags become available. Such 205–9. Bernatchez, L. and Wilson, C.C. (1998) Comparative devices, combined with other techniques used to phylogeography of nearctic and palearctic fishes. study population movement such as genetics and Molecular Ecology 7, 431–52. otolith microchemistry, can only improve our un- Bertin, L. (1956) Eels: A Biological Study. London: derstanding of how and why fish migrate, and Cleaver Hume. where they go. Block, B.A., Dewar, H., Williams, T., Prince, E.D., Farwell, C. and Fudge, D. (1998) Archival tagging of Atlantic bluefin tuna (Thunnus thynnus thynnus). Marine Technology Society Journal 32, 37–46. REFERENCES Bramwell, M. (ed.) (1977) The Mitchell Beazley Atlas of the Oceans. London: Mitchell Beazley. Allibone, R.M. and Wallis, G.P. (1993) Genetic variation Buckley, A.A. and Arnold, G.P. (2001) Orientation and and diadromy in some native New Zealand galaxiids swimming speed of plaice migrating by selective tidal (Teleostei: Galaxiidae). Biological Journal of the stream transport. In: J. Sibert and J. Nielsen (eds) Linnean Society 50, 19–33. Electronic Tagging and Tracking In Marine Fisheries. Arnold, G.P. (1981) Movement in relation to water Dordrecht: Kluwer Academic Press, pp. 263–77. currents. In: D.J. Aidley (ed.) Animal Migration. Carey, F.G. and Scharold, J.V. (1990) Movements of blue Cambridge, Cambridge University Press, pp. 55–79. sharks (Prionace glauca) in depth and course. Marine Arnold, G.P. and Cook, P.H. (1984) Fish migration by se- Biology 106, 329–42. lective tidal stream transport: first results with a com- Carre, F. and Geistdoerfer, P. (1981) Biologie et peche du puter simulation model for the European continental merlan bleu de l’Atlantique Nord. Peche Maritime 60, shelf. In: J.D. McCleave, G.P. Arnold, J.J. Dodson and 520–6. W.H. Neill (eds) Mechanisms of Migration in Fishes. Carscadden, J.E. (1984) Capelin in the Northwest New York: Plenum Press, pp. 227–61. Atlantic. In: D.N. Nettleship, G.A. Sanger and P.F. Arnold, G.P. and Dewar, H. (2001) Electronic tags Springer (eds) Marine Birds: Their Feeding Ecology and in marine fisheries research: a 30-year perspective. In: Commercial Fisheries Relationships. Special publica- J. Sibert and J. Nielsen (eds) Electronic Tagging tion compiled by the Canadian Wildlife Service for the and Tracking in Marine Fisheries. Dordrecht: Kluwer Pacific Seabird Group, pp. 170–82. Academic Press, pp. 7–64. Castonguay, M. and Gilbert, D. (1995) Effects of tidal Arnold, G.P. and Holford, B.H. (1995) A computer simu- streams on migrating Atlantic mackerel, Scomber lation model for predicting rates and scales of move- scombrus L. ICES Journal of Marine Science 52, ment of demersal fish on the European continental 941–54. shelf. ICES Journal of Marine Science 52, 981–90. Creaser, E.P., Libby, D.A. and Speirs, G.D. (1984) Sea- Arnold, G.P. and Metcalfe, J.D. (1996) Seasonal migra- sonal movements of juvenile and adult herring, Clupea tions of plaice (Pleuronectes platessa) through the harengus L., tagged along the Maine coast. Journal of Dover Strait. Marine Biology 127, 151–60. Northwest Atlantic Fishery Science 5, 71–8. Arnold, G.P., Greer Walker, M., Emerson, L.S. and Cushing, D.H. (1990) Hydrographic containment of a 196 Chapter 8

spawning group of plaice in the Southern Bight of the Copenhagen: International Council for the Explo- North Sea. Marine Ecology Progress Series 58, 287–97. ration of the Sea. Dingle, H. (1996) Migration: the Biology of Life on the Hamatsu, T. and Yabuki, K. (1995) Spawning migration Move. New York: Oxford University Press. and spawning ground of walleye pollock Theragra Dittman, A.H. and Quinn, T.P. (1996) Homing in Pacific chalcogramma distributed along the Pacific coast of salmon: mechanisms and ecological basis. Journal of eastern Hokkaido. Bulletin of the Hokkaido National Experimental Biology 199, 83–91. Fisheries Research Institute, Hokusuiken Hokoku 59, Dragesund, O., Hamre, J. and Ulltang, O. (1980) Biology 31–41. and population dynamics of the Norwegian spring Harden Jones, F.R. (1968) Fish Migration. London: spawning herring. Rapports et Procés-verbaux des Arnold. Réunions Conseil International pour l’Exploration de Harden Jones, F.R. (1980) The nekton: production and la Mer 177, 43–71. migration patterns. In: R.S.K. Barnes and K.H. Mann Drake, J.A. (1990) The mechanics of community assem- (eds) Fundamentals of Aquatic Ecosystems. Oxford: bly and succession. Journal of Theoretical Biology 147, Blackwell Science, pp. 119–42. 213–33. Harden Jones, F.R., Greer Walker, M. and Arnold, G.P. Drake, J.A. (1992) Community assembly mechanics and (1978) Tactics of fish movement in relation to migra- the structure of an experimental species ensemble. tion strategy and water circulation. In: H. Charnock American Naturalist 137, 1–26. and Sir George Deacon (eds) Advances in Oceanogra- Eschmeyer, W.N. (ed.) (1998) Catalogue of Fishes. San phy. New York: Plenum Publishing, pp. 185–207. Francisco: California Academy of Sciences. Harden Jones, F.R., Arnold, G.P., Greer Walker, M. and FAO (1997) Yearbook of Fishery Statistics, Vol. 84. Rome: Scholes, P. (1979) Selective tidal stream transport and FAO. the migration of plaice Pleuronectes platessa L. in the Fitzsimons, J.M. and Nishimoto, R.T. (1995) Use of fish southern North Sea. Journal du Conseil. Conseil behaviour in assessing the effects of hurricane Iniki on International pour l’Exploration de la Mer 38, 331–7. the Hawaiian island of Kaua’i. Environmental Biology Harding, D., Nichols, J.H. and Tungate, D.S. (1978) The of Fishes 43, 39–50. spawning of plaice (Pleuronectes platessa) in the Gauldie, R.W. and Sharp, G.D. (1996) Skipjack velocity, southern North Sea and English Channel. Rapports et dwell time and migration. Fisheries Oceanography 5, Procés-verbaux des Réunions Conseil International 100–13. pour l’Exploration de la Mer 172, 102–13. Gerking, S.D. (1959) The restricted movements of fish Hardisty, M.W. and Potter, I.C. (eds) (1971) The Biology of populations. Biological Reviews 34, 221–42. Lampreys. London: Academic Press. Glova, G.J. (1995) The secret life of the lamprey. Water Hasler, A.D. (1971) Orientation and fish migration. In: and Atmosphere 3, 20–1. W.S. Hoar and D.S. Randall (eds) Fish Physiology, Vol. Greer Walker, M., Harden Jones, F.R. and Arnold, G.P. 6. London: Academic Press, pp. 429–510. (1978) The movements of plaice (Pleuronectes Healey, M.C. (1991) Life history of chinook salmon, On- platessa) tracked in the open sea. Journal du Conseil. corhynchus tshawytscha. In: C. Groot and L. Margolis Conseil International pour l’Exploration de la Mer 38, (eds) Pacific Salmon Life Histories. Vancouver: 58–86. University of British Columbia, pp. 311–93. Groot, C. and Margolis, L. (eds) (1991) Pacific Salmon Life Hefford, A.E. (1909) The proportionate distribution of the Histories. Vancouver: University of British Columbia sexes of plaice in the North Sea. Rapports et Procés- Press. verbaux des Réunions Conseil International pour Gross, M.G. (1972) Oceanography, 2nd edn. Columbia, l’Exploration de la Mer 11, 135–76. OH: Charles E. Merrill Publishing. Hefford, A.E. (1916) Report on the sexual differentiation Gross, M.R., Coleman, R.M. and McDowall, R.M. (1991) in the biology and distribution of the plaice in the Aquatic productivity and the evolution of diadromous North Sea. Fishery Investigation, London Series 2, fish migration. Science 239, 1291–3. 3(2). Gunn, J. (1994) Smart archival tag comes up trumps for Heincke, F. (1992) The Varieties of Herring, Part 2. tuna. Australian Fisheries 53, 10–11. Canadian Translations of Fisheries and Aquatic Gunn, J., Polacheck, T., Davis, T., Sherlock, M. and Sciences, No. 5580. Betlehem, A. (1994) The development and use of Hendry, A.P. and Quinn, T.P. (1997) Variation in adult life archival tags for studying the migration, behaviour and history morphology among Lake Washington sockeye physiology of southern bluefin tuna, with an assess- salmon (Oncorhynchus nerka) populations in relation ment of the potential for transfer of the technology to to habitat features and ancestral affinities. Canadian groundfish research. ICES CM 1994/Mini: 2.1 (21). Journal of Fisheries and Aquatic Sciences 54, 75–84. Migration 197

Hill, R.D. (1994) Theory of geolocation by light levels. In: try system for transmitting body and water tempera- B.J. LeBoeuf and R.M. Laws (eds) Elephant Seals: Popu- ture from free swimming fish. Woods Hole, MA: Tech- lation Ecology, Behavior, and Physiology. Berkeley, nical Report of the Woods Hole Oceanographic CA: University of California Press, pp. 227–36. Institute WHOI-71-67 (unpublished manuscript). Hill, R.D. and Braun, M.J. (2001) Geolocation by light Leider, S.A. (1989) Increased straying by adult steelhead levels – the next step: latitude. In: J. Sibert and J. trout, Salmo gairdneri, following the 1980 eruption of Nielsen (eds) Electronic Tagging and Tracking in Mount St. Helens. Environmental Biology of Fishes Marine Fisheries. Dordrecht: Kluwer Academic Press, 24, 219–29. pp. 315–30. Lin, X. (1987) Biological characteristics and resources Houghton, R.G. and Harding, D. (1976) The plaice of the status of three main commercial fishes in offshore English Channel: spawning and migration. Journal du waters of China. Journal of Fisheries of China 11, Conseil. Conseil International pour l’Exploration de 187–94. la Mer 36, 229–39. Lutcavage, M.E., Brill, R.W., Skomal, G.B. Chase, B.C. Hoziosky, S.A., Shvetsov, F.G. and Gradalev, E. (1989) and Howey, P.W. (1999) Results of pop-up satellite tag- Seasonal distribution, migration, and mortality com- ging of spawning size class fish in the Gulf of Maine: ponent dynamics in sprat of the eastern and southeast- do North Atlantic bluefin tuna spawn in the mid ern Baltic. Rapports et Procés-verbaux des Réunions Atlantic? Canadian Journal of Fisheries and Aquatic Conseil International pour l’Exploration de la Mer Sciences 56, 173–7. 190, 22–4. MacCrimmon, H.R. and Gots, B.L. (1979) World distribu- Hunter, J.R., Argue, A.W., Bayliff, W.H. et al. (1986) The tion of Atlantic salmon, Salmo salar. Journal of the dynamics of tuna movements: an evaluation of past Fisheries Research Board of Canada 36, 422–57. and future research. FAO Fisheries Technical Paper MacLennan, D.N. and Simmonds, E.J. (1991) Fisheries 277, 1–78. Acoustics. London: Chapman & Hall. Hutchinson, W.F., Carvalho, G.R. and Rogers, S.I. (2001) Main, M.R. (1989) Distribution and post glacial dispersal Marked genetic structuring in localised spawning of freshwater fishes in South Westland, New Zealand. populations of cod Gadus morhua in the North Sea Journal of the Royal Society of New Zealand 19, 161–9. and adjoining waters, as revealed by microsatellites. Malmberg, S.A. (1980) Hydrographic conditions and fish Marine Ecology Progress Series 223, 251–60. stocks in Icelandic waters. 4. Hydrographic conditions ICCAT (1999) Report for the Biannial Period, 1998–1999, and migration of cod. Aegir 73, 205–11. Part I, p. 54. International Council for the Conserva- McCleave, J.D. and Arnold, G.P. (1999) Movements of tion of Atlantic Tuna. yellow- and silver-phase European eels (Anguilla Jamieson, A. and Thompson, D. (1972) Blood proteins in anguilla L.) tracked in the western North Sea. ICES North Sea cod (Gadus morhua L.) In: G. Kovacs and M. Journal of Marine Science 56, 510–36. Papp (eds) Proceedings XII European Conference on McDowall, R.M. (1988) Diadromy in Fishes: Migrations Animal Blood Groups and Biochemical Polymor- Between Freshwater and Marine Environments. phism. The Hague: Dr W. Junk, pp. 585–91. London: Croom Helm. Jellyman, D.J. (1995) Longevity of longfinned eels, An- McDowall, R.M. (1990) New Zealand Freshwater Fishes: guilla dieffenbachii, in a New Zealand high country A Natural History and Guide. Auckland: Heinemann lake. Ecology of Freshwater Fish 4, 106–12. Reed. Jennings, S. and Pawson, M.G. (1991) The development McDowall, R.M. (1993) Implications of diadromy for the of bass, Dicentrarchus labrax, eggs in relation to tem- structuring and modelling of riverine fish communi- perature. Journal of the Marine Biological Association ties in New Zealand. New Zealand Journal of Marine of the UK 71, 107–16. and Freshwater Research 27, 453–62. Jennings, S., Kaiser, M.J. and Reynolds, J.D. (2001) McDowall, R.M. (1995) Understanding the absence of Marine Fisheries Ecology. Oxford: Blackwell Science. upland bullies from eastern Banks Peninsula. Water Karr, J.R. (1991) Ecological integrity: a long neglected and Atmosphere 3, 17–19. aspect of water resource management. Ecological Ap- McDowall, R.M. (1996) Volcanism and freshwater fish plications 1, 66–84. biogeography in the northeastern North Island of New Kawanabe, H. (1969) The significance of social structure Zealand. Journal of Biogeography 23, 139–48. in production of the ‘ayu’, Plecoglossus altivelis. In: McDowall, R.M. (1998a) Driven by diadromy: its role in T.G. Northcote (ed.) Symposium on Salmon and Trout the historical and ecological biogeography of New in Streams. Vancouver: University of British Colum- Zealand freshwater fishes. Italian Journal of Zoology bia, pp. 243–51. 65 (Suppl. S), 73–85. Lawson, K.D. and Carey, F.G. (1972) An acoustic teleme- McDowall, R.M. (1998b) Fighting the flow: downstream 198 Chapter 8

upstream linkages in the ecology of diadromous fish migratory behaviour of plaice in the North Sea as re- faunas in West Coast New Zealand rivers. Freshwater vealed by data storage tags. CM 1994/Mini:11. Copen- Biology 40, 111–22. hagen: International Council for the Exploration of the McDowall, R.M. (1999) Diadromy and genetic diversity Seas. in Nearctic and Palearctic fishes. Molecular Ecology 8, Metcalfe, J.D., Arnold, G.P. and Hunter, E. (1998) The use 527–8. of electronic data storage tags in modelling the effec- McDowall, R.M. and Taylor, M.J. (2000) Environmental tiveness of marine protected areas. ICES CM 1998/U:4. indicators of habitat quality in a migratory fresh- Copenhagen: International Council for the Explo- water fish fauna. Environmental Management 25, ration of the Sea. 357–74. Metcalfe, J.D., Hunter, E., Holford, B.H. and Arnold, G.P. McGladdery, S.E. and Burt, M.D.B. (1985) Potential of (1999) The application of electronic data storage tags parasites for use as biological indicators of migration, to spatial dynamics of fish populations. ICES CM feeding, and spawning behavior of northwestern At- 1999/AA:01. Copenhagen: International Council for lantic herring (Clupea harengus). Canadian Journal of the Exploration of the Sea. Fisheries and Aquatic Sciences 42, 1957–68. Milner, A.M. (1987) Colonization and ecological devel- McKeown, B.A. (1984) Fish Migration. London: Croom opment of new streams in Glacier Bay National Park, Helm. Alaska. Freshwater Biology 18, 53–70. McKnight, D.M. and Dahm, C.N. (1990) Contribution of Milner, A.M. and Bailey, R.G. (1989) Salmonid coloniza- organic acids to alkalinity in lakes within the Mount tion of new streams in Glacier Bay National Park, St. Helens blast zone. Limnology and Oceanography Alaska. Aquaculture and Fisheries Management 20, 35, 535–42. 179–92. McQuinn, I.H. (1997) Metapopulations and the Atlantic Minshall, G.W., Cummins, K.W., Petersen, R.C. et al. herring. Reviews in Fish Biology and Fisheries 7, (1985) Developments in stream ecosystem theory. 297–329. Canadian Journal of Fisheries and Aquatic Sciences Meek, A. (1916) The Migrations of Fishes. London: 42, 1045–55. Arnold. Moore, R. (1982) Spawning and early life history of Metcalfe, J.D. (2001) Summary report of a workshop on barramundi, Lates calcarifer (Bloch) in Papua New daylight measurements for geolocation in animal Guinea. Australian Journal of Marine and Freshwater telemetry. In: J. Sibert and J. Nielsen (eds) Electronic Research 33, 647–61. Tagging and Tracking In Marine Fisheries. Dordrecht: Moore, R. and Reynolds, L.F. (1982) Migration patterns of Kluwer Academic Press, pp. 331–42. barramundi, Lates calcarifer (Bloch) in Papua New Metcalfe, J.D. and Arnold, G.P. (1997) Tracking fish with Guinea. Australian Journal of Marine and Freshwater electronic tags. Nature 387, 665–6. Research 33, 671–82. Metcalfe, J.D. and Arnold, G.P. (1998) Tracking migra- Myers, G.S. (1949) Usage of anadromous, catadromous tion fish with electronic tags. In: M. McCarthy (ed.) and allied terms for migratory fishes. Copeia 1949, EEZ Technology: A Review of Advanced Technologies 89–97. for the Management of EEZs World Wide, 2nd edn. Nelson, D.R. (1978) Telemetering techniques for the London: ICG Publishing, pp. 199–206. study of free-ranging sharks. In: E.S. Hodgson and R.F. Metcalfe, J.D., Arnold, G.P. and Webb, P.W. (1990) The Mathewson (eds) Sensory Biology of Sharks, Skates energetics of migration by selective tidal stream trans- and Rays. Arlington, VA: Office of Naval Research, port: an analysis for plaice tracked in the southern Department of the Navy, pp. 419–82. North Sea. Journal of the Marine Biological Associa- Novikov, Yu.V. (1986) Some trends of distribution and tion of the UK 70, 149–62. migration of common pelagic fishes of the north- Metcalfe, J.D., Fulcher, M. and Storeton West, T.J. (1992) western Pacific Ocean. Journal of Ichthyology 26, Progress and developments in telemetry for monitor- 161–72. ing the migratory behaviour of plaice in the North Sea. Okada, Y. (1960) Studies on the Freshwater Fishes of In: I.G. Priede and S.M. Swift (eds) Wildlife Telemetry: Japan. Tsu, Japan: Prefectural University of Mie. Remote Monitoring and Tracking of Animals. London: Parker, N.C., Giorgi, A.E., Heidinger, R.C., Jester D.B., Ellis Horwood, pp. 359–66. Prince, E.D. and Winans, G.A. (eds) (1990) Fish- Metcalfe, J.D., Arnold, G.P. and Holford, B.H. (1993) Ori- marking Techniques. Bethesda, MD: American Fish- entation of plaice (Pleuronectes platessa) in the open eries Society. sea: evidence for the use of external directional clues. Pawson, M.G. and Pickett, G.D. (1996) The annual pat- Marine Biology 117, 559–66. tern of condition and maturity in bass, Dicentrarchus Metcalfe, J.D., Arnold, G.P. and Holford, B.H. (1994) The labrax, in waters around England and Wales. Journal of Migration 199

the Marine Biological Association of the UK 76, Templeton, R.G. and Churchward, A. (1990) Fisheries 107–25. management practices of the Severn Trent Water Pawson, M.G., Kelley, D.F. and Pickett, G.D. (1987) The Authority in England 1976–1986. In: Densen, Tvan, B. distribution and migrations of bass, Dicentrarchus Steinmetz and R.H. Hughes (eds) Management of labrax L., in waters around England and Wales as Freshwater Fisheries. Proceedings of a Symposium shown by tagging. Journal of the Marine Biological Organised by the European Inland Fisheries Advisory Association of the UK 67, 183–217. Commission. Wageningen, Netherlands: Pudoc, pp. Pawson, M.G., Pickett, G.D. and Witthames, P.R. (2000) 558–68. The influence of temperature on the onset of first Tesch, F.W. (1977) The Eel. Biology and Management of maturity in sea-bass. Journal of Fish Biology 56, Anguillid Eels. London: Chapman & Hall. 319–27. Thompson, B.M. and Harrop, R.T. (1987) The distribu- Pickett, G.D. and Pawson, M.G. (1994) Sea Bass: Biology, tion and abundance of bass (Dicentrarchus labrax) eggs Exploitation and Management. London: Chapman & and larvae in the English Channel and southern North Hall. Sea. Journal of the Marine Biological Association of Punt, A.E. and Butterworth, D.S. (1994) Use of tagging the UK 67, 263–74. data within a VPA formalism to estimate migration Tosello Bancal, F. (1989) Apercu historique de la peche de rates of bluefin tuna across the North Atlantic. Collec- la sardine sur les cotes atlantiques francaises. Peche tive Volume of Scientific Papers. International Com- Maritime 68, 701–5. mission for the Conservation of Atlantic Tunas 44, Trout, G.C. (1957) The Bear Island Cod: Migrations and 166–82. Movements. Fishery Investigations, London 21(6). Rey, J.C. (1983) El paso del atún rojo, Thunnus thynnus Tsukamota, K. (1992) Discovery of the spawning area for (Linnaeus, 1758), a través del Estrecho de Gibraltat y su Japanese eel. Nature 356, 789–91. relación con la hidrología. Esquemas de magración. Vannote, R.L., Minshall, G.W., Cummins, K.W., Sedell, Boletin del Instituto Espanol de Oceanografia 1, J.R. and Cushing, C.E. (1980) The river continuum 85–94. concept. Canadian Journal of Fisheries and Aquatic Righton, D., Metcalfe, J. and Connolly, P. (2001) Different Sciences 37, 130–7. behaviour of North and Irish Sea cod. Nature 411, 156. Walker, R.V., Myers, K.W., Davis, N.D. et al. (2000) Diur- Roberts, T.R. (1978) An ichthyological survey of the Fly nal variation in the thermal environment experienced River in Papua New Guinea with descriptions of new by salmonids in the north Pacific as indicated by data species. Smithsonian Contributions to Zoology 281, storage tags. Fisheries Oceanography 9, 171–86. 1–72. Walton, I. and Cotton, C. (1898) The Complete Angler, or Rose, G.A. (1993) Cod spawning on a migration highway the contemplative man’s recreation. Boston: Little, in the North west Atlantic. Nature 366, 458–61. Brown. Schmidt, J. (1922) The breeding places of the eel. Philo- Watanabe, Y., Ogura, M. and Tanabe, T. (1995) Migration sophical Transactions of the Royal Society of London of skipjack tuna, Katsuwonus pelamis, in the western Series B 211, 179–208. Pacific Ocean, as estimated from tagging data. Bulletin Schmidt, J. (1923) Breeding places and migrations of the of the Tohoku National Fisheries Research Institute eel. Nature 111, 51–4. Tohokusuikenho 57, 31–60. Scott, W.B. and Crossman, E.J. (1973) Freshwater fishes of Weihs, D. (1978) Tidal stream transport as an efficient Canada. Bulletin of the Fisheries Research Board of method for migration. Journal du Conseil 38, Canada 184, 1–996. 92–9. Serra, R. (1991) Important life history aspects of the Wimpenny, R.S. (1953) The Plaice. London: Arnold. Chilean jack mackerel, Trachurus symmetricus mur- Woodhead, A.D. (1959) Variations in the activity of the phyi. Investigation Pesquara 36, 67–83. thyroid gland of the cod, Gadus callarias L., in relation Simpson, A.C. (1959) The spawning of the plaice (Pleu- to its migrations in the Barents Sea. II. The ‘dummy ronectes platessa) in the North Sea. Fishery Investiga- run’ of immature fish. Journal of the Marine Biological tions, London 22(7). Association of the UK 38, 417–22. Slotte, A. and Fiksen, Ø. (2000) State dependent spawn- Wu, G. (1989) The ecological characteristics of distribu- ing migration in Norwegian spring spawning herring. tion of eggs, larvae and juveniles of the Engraulis Journal of Fish Biology 56, 138–62. japonicus (Temminck and Schlegel) and Anchoviella Suzuki, Z., Tomlinson, P.K. and Honma, M. (1978) Popu- commersonii (Lecepede) in the Changjiang River lation structure of Pacific yellowfin tuna. Bull. IATTC Estuary. Oceanologia et Limnologia Sinica 20, Bol. CIAT 17, 277–441. 217–29. 9 Genetics of Fish Populations

ROBERT D. WARD

9.1 INTRODUCTION to define stock structures of commercial fishes; they are also important in other areas of concern The well-publicized collapses of many well- to fisheries managers, including conservation and known fish stocks, such as those of Peruvian enhancement programmes and the validation of anchoveta (Engraulis ringens), North Sea herring species identification. In this chapter, the genetic (Clupea harengus) and Newfoundland cod (Gadus tools currently available are described, with their morhua) (Beverton 1990; Hutchings and Myers strengths and weaknesses, and examples given of 1994), have led to demands for more effective and their uses in specimen identification, population more sustainable management. Fishery managers structure analyses and conservation and enhance- are now primarily aiming for resource sustain- ment programmes. The chapter finishes with a ability and the recovery of severely depleted short section on the genetics of sex determination populations. Their decisions are based on stock in fish, and some concluding remarks. assessments, together with socioeconomic and sometimes political considerations. Stock identi- fication is an integral element (Carvalho and 9.2 GENETIC TOOLS Hauser 1994; Begg et al. 1999); indeed a recent check-list of stock assessment methods (Deriso Genetic analysis of fish populations is not new. and Quinn 1998) listed information on spatial dis- Serological approaches were first adopted in the tribution or stock structure as the number one early 1930s (see de Ligny 1969) but failed to gain item. widespread acceptance. The genetic tools now Determining the stock or population structure used are based on electrophoretic analysis of of any fish species is a complex task. There are proteins and DNA. Table 9.1 summarizes their many approaches that can be taken, including properties, albeit in a somewhat simplified and genetic analyses, phenotype analyses to detail subjective fashion, and should be read in conjunc- growth rates, age composition, morphometrics tion with the following subsections. Table 9.2 and microconstituents in calcified structures, to- summarizes their uses in different areas of fish- gether with parasite loads and tagging returns. eries science. While these different approaches generally com- plement each other and assist in a better under- 9.2.1 Allozyme analysis standing of stock structure, their reconciliation into a unified model can be challenging (McQuinn The first genetic tool to be widely adopted for fish 1997). population analysis was protein electrophoresis. Genetic analyses are not only useful in helping One of the earliest studies was Sick’s (1961) Genetics 201 ssarily tissue available tissue conditions comparisons mer sequences are ce is summarized in preservedpreservedpopular not blots; potential inherited preserved inherited preserved based required markersmarkers of non-coding detectable proteindetectable variation following variation enzyme restriction digestion maternally inherited frozen frozen,of repeats tandem of preserved dinucleotide/motifstetranucleotide longer repeats codingmarker single (some alleles)suitable if tool null recessive frozen, preserved frozen,deterrent; problems a scoring unknown not Southern with Basis PCR Speed Expense Genetics Tissueof Number Variability Coding/ Comments The major current genetic tools used in fisheries science, with some general comments on their properties. The comments are nece partly subjective and sometimes oversimplified. For example, the costs of microsatellite analysis are significantly less when pri Table 9.1 Table The suitability of these tools in different areas fisheries scien already available than when they have to be developed anew. 9.2. Table Allozymes ElectrophoreticallyMtDNA-RFLP Nosize Fragment MtDNA-PCR- *****RFLP No *usingbut above, As MtDNA Yessequencing ** Codominantproducts PCR Fresh, ***Randomly variationsequence Nucleotide ***amplified *** Usuallypolymorphic *** Haploid,sizefragment (RAPD) DNA DNA ***offollowingprimers variation of use Microsatellite Haploid, ** Fresh, Yessequence random ****numberin Variation Fresh, * Haploid, Yes Coding ****Minisatellite * ** Fresh,excellent an Still tandembut ** above As *** ** SometimesAmplified Dominant ** ***lengthfragment *** maternallyafterfragments polymorphism of Mostly Fresh, frozen, ***(AFLP) Presence/absence Codominant ** selective maternally Mostly *****Exon-primed Fresh,a Effectively Yes Codominantintron-crossing frozen,(EPIC) PCR sizeor Sequence Fresh, *** *****intronsa in Effectively variation Mostly amplification restriction fragment length polymorphism; mtDNA, mitochondrial DNA. PCR, polymerase chain reaction; RFLP, **** *** ***** Yesa Effectively Mostly frozen, ** preserved Non-coding ***a time Reproducibility Development *** Dominant coding ** Fresh, Non-codingmarker scored single Usually coding ***** Codominant Fresh,marker single *** frozen, non-coding ***highly concern; Mostly ***?fish; for novel PCR Still preservedto sensitive frozen, Non-codingbut novel, Still non-codingcan bands multiple cross-gel hinder considerable with 202 Chapter 9

Table 9.2 Suitability/uses of the current genetic tools in fisheries science. Columns graded * to *****; for example allozymes are very suitable for species identification and somewhat less so for aquaculture genetics. Scorings are overall scores; for example mtDNA sequencing is a very powerful approach to species identification but given its expense is rated below the mtDNA-PCR-RFLP approach. Scores marked with a ? indicate that these approaches have been little explored as yet, but that these are likely scores based on the properties of these markers. Column headings are broad; for example, aquaculture genetics embraces diversity estimation, pedigree analysis and gene mapping studies. Some tools might be better in some subcategories than others; for example AFLPs will likely be of more use in aquaculture in mapping than studies of gene diversity.

Species identification Population genetics Systematics Aquaculture genetics

Allozymes ***** *** *** *** MtDNA-RFLP **** ** *** * MtDNA-PCR-RFLP ***** *** *** ** MtDNA sequencing **** ** ***** * RAPD *** ** * ** Microsatellite * **** * ***** Minisatellite * *** * ** AFLP ***? ** *? *** EPIC **? ****? **? ***?

AFLP, amplified fragment length polymorphism; EPIC, exon-primed intron-crossing PCR; mtDNA, mitochondrial DNA; PCR, polymerase chain reaction; RAPD, randomly amplified polymorphic DNA; RFLP, restriction fragment length polymorphism.

examination of haemoglobin variation in whiting with the limited number of loci that can be as- (Merlangius merlangus) and cod (Gadus morhua). sessed (usually <50), can give little variation for The subsequent advent of enzyme-specific histo- screening purposes. There is also a lingering debate chemical staining made it possible to examine a over whether the variation detected is or is not large range of enzymes, a procedure often referred selectively neutral. Lewontin (1974) gives a classic to as allozyme analysis. The seminal studies of account of this controversy. Population genetic Lewontin and Hubby (1966) and Harris (1966) interpretation is generally based on the null showed that genetically determined allozyme hypothesis of selective neutrality, so that popula- variation was widespread in Drosophila and tion differences are held to reflect reproductive iso- human populations. Fish populations were soon lation and genetic drift. However, it is possible that shown to be no different, and for the first time a some of the amino acid differences responsible for simple, quick and reliable genetic method was allozyme separation affect function in some way; available for examining fish population structure. if this is so, then population differences might Practical details may be found in Richardson et al. reflect differences in selection pressures rather (1986), Whitmore (1990), Manchenko (1994) and than isolation. Hillis et al. (1996). A modification of protein electrophoresis is Allozyme analysis remains a highly attractive isoelectric focusing (see Whitmore 1990), where approach, when fresh or frozen tissue is available, proteins are electrophoretically separated on a pH in terms of speed and cost. Many allozymes are gradient gel on the basis of their surface electrical tissue-specific, so the availability of different tis- charge alone, rather than size and charge as in sues including muscle, liver and nerve (eye) is ad- conventional electrophoresis. They are sharply vantageous. Significant disadvantages include low ‘focused’ at their isoelectric point, the pH at which average degrees of variability which, when coupled the protein has a net surface charge of zero. Genetics 203

9.2.2 DNA analysis Male Female a c The ability to examine DNA variation directly, Nuclear DNA b rather than indirectly through protein variants, d came with the recognition and isolation of restric- Mitochondrial X Y tion enzymes. There are many such enzymes, DNA which cut or ‘restrict’ DNA at enzyme-specific nucleotide sequences, usually of four or six bases. Unfortunately, with the constant development of Genotypes of offspring (male or female) new analytical DNA techniques, comprehensive manuals are lacking. Two good sources of in- formation are Hillis et al. (1996), who provide a abb an overview of molecular techniques and data c d c d analysis, and Caetano-Anollés and Gresshoff (1997), who provide details on selected methods of Y YY Y DNA marker analysis and their applications, in- cluding methods not discussed here. A brief out- Fig. 9.1 Inheritance of nuclear DNA and mitochondrial line of some of the major and more commonly used DNA. The male parent has genotype a/b at a nuclear DNA locus, and mitochondrial haplotype X. The female methods in fisheries population genetics follows; parent has genotype c/d at the nuclear DNA locus, and other approaches are included, briefly, in Tables mitochondrial haplotype Y. The offspring have one of 9.1 and 9.2. four nuclear DNA genotypes (a/c, a/d, b/c and b/d) but all have the mitochondrial DNA haplotype Y. Mitochondrial DNA Most initial DNA analyses were of mitochondrial DNA (mtDNA), largely because this genome ing it possible to choose a relatively variable region (usually about 16–18kb) is much smaller than for population studies or a relatively conserved nuclear DNA (nDNA) and was easier to examine region for taxonomic or systematic studies. with the techniques available some 10–15 years However, mtDNA also has significant disad- ago. Mitochondrial DNA was extracted and cut vantages. Principal among these is that it is a non- with restriction enzymes; the resulting fragments recombining genome and best treated as a single were separated by electrophoresis and visualized. character, whereas nDNA assessment can be This is a restriction fragment length polymor- based on many independent characters (loci). phism (RFLP) analysis. Mitochondrial DNA is inherited differently Polymerase chain reaction from nDNA (Fig. 9.1): it is only inherited from the maternal parent, with very rare exceptions, and is Analyses of DNA variation, both mtDNA and haploid in nature. The genetically effective popu- nDNA, were considerably enhanced by the lation size of mtDNA is thus, other factors being development of polymerase chain reaction (PCR) equal, only one-quarter that of nDNA, which is techniques. PCR analysis allows the repeated inherited from both parents and is diploid in each amplification of a DNA region lying between parent. This makes it potentially a more sensitive two DNA primers, the primers generally being indicator of genetic drift than nDNA. Further- sequences of around 20bp that are complementary more, its overall rate of evolution is about an order to sequences in the target DNA. These techniques of magnitude faster than single-copy nDNA. The typically allow the amplification of a DNA region, non-coding control region (d-loop) of mtDNA flanked by the complementary primer sequences, evolves more rapidly than coding segments, mak- that is some hundreds of base pairs long, although 204 Chapter 9 regions of several thousand base pairs can now be labels and an automated DNA sequencer, al- amplified successfully. Once the target fragment though non-automated approaches are also has been amplified, it can be examined for size or popular. Unfortunately, primers tend to be specific sequence variation. to an individual species or species group, as muta- This general approach is very flexible and lends tions that accumulate in the flanking region itself to a wide variety of adaptations and modifica- quickly reduce the binding ability of the primers. tions. Many of these relate to primer design Microsatellite primers generally have to be devel- and make it possible to target different classes oped anew for each species or species group; this of mtDNA or nDNA sequence. The amplified time-consuming development phase can require fragments can be assessed by a wide variety of several months of skilled labour. methods, including RFLP analysis, sequencing, Microsatellites provide an abundant supply of size analysis, single-strand conformational poly- hypervariable codominant markers for fish studies morphisms and dot plots. Furthermore, very little (O’Connell and Wright 1997). This is an advantage tissue is needed as the target DNA is amplified in pedigree or mapping studies, but too many very many-fold from the small amount initially avail- low frequency alleles can reduce the power of able. This permits non-invasive and non-lethal population structure analyses. Scoring of geno- sampling. DNA can be extracted and amplified types can sometimes prove difficult, especially from old fish scales (Nielsen et al. 1999) or when there are many alleles of similar size and museum specimens (Pichler and Baker 2000), when PCR products of a single allele produce permitting historic surveys of changes in genetic multiple bands on gels, a phenomenon known as variability. Tissue can be stored at room tempera- band ‘stuttering’ or ‘laddering’. Non-amplifying or ture in alcohol or other chemicals such as ‘null’ alleles have also been recorded; these can dimethyl sulfoxide, simplifying the logistics of arise from mutations in the primer site so that a sampling and freight. PCR analyses have now particular allele is not amplified or from the PCR largely supplanted non-PCR analyses. process not amplifying a long allele as well as a short allele. Null alleles are recessive alleles. In population surveys, an individual heterozygous Microsatellites for an expressed allele and a null allele will usually Microsatellites are nDNA regions of repeated be mistakenly scored as homozygous for the ex- sequences where the repeat sequence or motif is pressed allele. short, 2–5bp. An example is TCTCTCTC, which is a 2-bp or dinucleotide repeat. The total micro- Randomly amplified polymorphic DNA satellite length is usually less than 300bp (Tautz 1989). Microsatellites are flanked by non- Another PCR technique, less widely used in fish- repeat sequences that are conserved within eries, is the analysis of randomly amplified poly- species, making it possible to design PCR primers morphic DNA (RAPD; Welsh and McClelland which will amplify the microsatellite region. 1990; Williams et al. 1990). Primers of random Microsatellite variability arises principally from nucleotide sequence, typically around 10–20bp, variation in the number of repeats of the motif, are used to amplify anonymous segments of DNA. leading to size changes in the PCR product that can Variation in primer binding sites leads to variation be identified on electrophoresis gels. Microsatel- in PCR products, which are again scored from size lites are usually in non-coding regions and thus variation in electrophoresis gels. This is a quick likely to be selectively neutral. The mutation and relatively simple technique, although amplifi- rate of microsatellite loci is very high, and cation occurs in conditions of low stringency consequently dozens of alleles per microsatellite and repeatability of observations requires careful locus are not uncommon. Allele data are most attention to all aspects of the procedure. It is a readily collected using primers with fluorescent dominant/recessive system, and DNA fragments Genetics 205 of any particular locus in any individual are gener- of distantly related species (e.g. Quattro and Jones ally scored as present (genotypes + + or + -) or ab- 1999). sent (genotype - -). 9.2.3 Uses Amplified fragment length polymorphism A somewhat subjective attempt to summarize the Amplified fragment length polymorphism (AFLP; abilities of these different tools in four areas of fish- Vos et al. 1995) constitutes a third major marker eries science is provided in Table 9.2. Allozymes class. In this procedure, genomic DNA is digested and mtDNA analysis are excellent tools for species with two restriction enzymes, EcoRI and MseI, identification. Allozymes and microsatellites and suitable adapters ligated on to the fragments. are perhaps the best tools for population genetics, Primers are then used that recognize these although neither rates the full five-star category. adapters and also have additional terminal selec- Allozymes have limited variability and lingering tive bases, which can be used to discriminate uncertainties concerning the relative roles of between and then amplify particular subsets of selection and drift, whilst microsatellites can fragments. By changing these selective bases, dif- have too much variability and scoring difficulties. ferent sets of fragments can be recognized. Like the Systematics and aquaculture are important RAPD approach it does not require prior sequence areas of fish genetics not otherwise covered here. information, but unlike the RAPD approach Mitochondrial DNA sequencing is perhaps the primer specificity is high and stringent annealing best current tool for systematics but may in future conditions can be used. Repeatability is therefore be equalled or bettered by nDNA sequencing. higher. It can produce large numbers of poly- Microsatellites are probably the best overall tool morphic markers on a single gel. One disadvantage for aquaculture applications and would be used to of AFLP is that, like RAPD, it is generally scored provide measures of diversity in hatchery stocks, as a dominant/recessive system. Mueller and to assess pedigrees and to map genomes. AFLPs Wolfenbarger (1999) give a short general account yield large numbers of markers very quickly for of the methods and uses of AFLP. mapping purposes.

Exon-primed intron-crossing PCR 9.3 STATISTICAL TOOLS A further PCR variant makes use of the fact that many genes in higher organisms are made up of The development of hypervariable markers has exon and intron sequences. Exons are expressed rendered traditional statistical methods for whilst introns are intervening sequences that do analysing genotype and allele frequencies inappro- not code for any product. Generally the exon re- priate or at least lacking in power. Traditional gions, which code for amino acids in protein genes, approaches cannot deal with very small cell or are highly conserved, but the intervening non- sample sizes, and require the pooling of rare geno- coding sequences can accumulate mutations. types or alleles. The new methods typically use Thus conserved primers that anneal to adjacent computer-intensive resampling techniques that exon regions can be created which amplify the do not require data pooling (Rousset and Raymond variable DNA from introns; this is exon-primed 1997). Many of the statistical packages for intron-crossing PCR (EPIC). Intron variation can analysing molecular data are briefly described in be assessed by RFLP analysis, fragment size elec- Hillis et al. (1996) and Schnabel et al. (1998). Most trophoresis or any other appropriate technique. implement Monte Carlo or other methods for EPIC analysis has the advantage that a particular dealing with hypervariable loci and small cell pair of exon primers may be relatively universal. sizes. Maximum likelihood and coalescence This means that they are usable over a wide range methods are being explored as alternatives to tradi- 206 Chapter 9 tional ways of handling gene frequencies and evolutionary rate so it is difficult to find such a estimating gene flow rates (e.g. Neigel 1997; Beerli marker. However, some markers show little 1998; Luikart and England 1999). variation within species but enough variation be- tween species to make them suitable for species identification. Allozyme analysis is a satisfactory method of 9.4 SPECIMEN AND SPECIES identifying fish tissue, provided that the material IDENTIFICATION is either fresh or frozen. The methods are suffi- ciently sensitive to be able to identify fish eggs and Accurate species identification is vital for scien- larvae (see Ward and Grewe 1994 for examples). A tific research, for catch recording and for labelling fast and simple method for fish identification in- in the marketplace. Accurate catch identification volves Coomassie blue staining of the major white is necessary to monitor species managed under muscle proteins after electrophoresis (Shaklee and quota systems; some species are morphologically Keenan 1986). The use of a cellulose acetate matrix very similar yet under quite different management rather than starch or acrylamide can speed the regimes. Accurate labelling is necessary for con- time required to less than 1h. A recently published sumer confidence in fish products. Cases of fish identification guide for Australia’s domestic substitution or mislabelling do occur, more com- seafood includes cellulose acetate-based protein monly where fish are marketed as fillets rather fingerprints of about 380 species (Yearsley et al. than as whole fish. Compliance officers require a 1999). Where possible, multiple specimens of tool that provides unequivocal specimen identifi- each species were examined so that at least the cation and which will be acceptable to a court of more common intraspecies variants were detected law. Finally, eggs and larvae may be virtually im- where present. This simple test cannot differen- possible to identify morphologically, yet accurate tiate all members of some closely related species identification may be required for research or as- groups; in such instances supplementary allozyme sessment purposes. or DNA tests are required. Molecular genetic methods have proven in- Isoelectric focusing of muscle proteins has valuable in all these areas, and indeed have some- been used to identify red snapper (Lutjanus times led to the recognition of previously campechanus) fillets (Huang et al. 1995) and caviar unrecognized sibling species (see Ward and Grewe from various sturgeon species (Chen et al. 1996), 1994 for examples). The premise underlying and can be used to discriminate snapper cooked genetic identification of species is straightforward: in a variety of ways (Hsieh et al. 1997). The cooked if two taxa are reproductively isolated from each flesh of ten commercially important species was other and warrant species status, they are likely to examined in nine laboratories using the two elec- have accumulated sufficient genetic differences trophoretic methods SDS polyacrylamide and urea during their isolation to be genetically distinguish- isoelectric focusing. With the exception of some able. If the two taxa are sympatric, genetic dif- closely related salmonids, all species could be ferentiation provides strong evidence of species identified (Etienne et al. 2000). Isoelectric focusing separation; if they are allopatric, then large genetic methods have also been used in the US-based differences again suggest two discrete species, but Regulatory Fish Encyclopedia (Office of Seafood small differences could be compatible with either and Office of Regulatory Affairs, Food and Drug two recently diverged species or limited gene flow Administration, 1993–2000, see the website: between populations of a single species. http://vm.cfsan.fda.gov/~frf/rfe0.html). The method chosen should ideally target DNA-based methods can be used when speci- a genetic marker of low intraspecies variability mens cannot be identified by protein or allozyme but high interspecies variability. Contemporary electrophoresis. Examples would be when species levels of variation generally correlate with are genetically so similar that diagnostic proteins Genetics 207 cannot be found, when insufficient tissue is (Paralithodes camtschatica) to an area closed to available or when specimens have been preserved fishing rather than to the claimed open area. in such a way as to destroy protein integrity or One point worth making here is that for many enzyme activity. Early cases of fish identification population genetic surveys, tissue samples are using PCR amplification of mtDNA, followed by collected by third parties or relatively untrained sequencing or RFLP analysis, are given in Ward and personnel. In such situations checks on the species Grewe (1994). More recent examples are billfish identity of the tissue fragments prior to population (Chow 1994; Innes et al. 1998), tunas (Ward et al. analysis are highly desirable. This can be done 1995), flatfish (Cespedes et al. 1998), fish eggs quickly by mtDNA examination, which might (caviar) (DeSalle and Birstein 1996) and starfish also provide data on population structure to com- larvae (Evans et al. 1998). Rehbein et al. (1999) plement any nDNA analysis. used single-strand conformation polymorphism of mtDNA to identify a variety of raw and canned fishery products. 9.4.1 Identifying and studying Nuclear DNA identification is less common, fish hybridization although RAPD has been used for tilapias (Dinesh Hybridization between fish species is not uncom- et al. 1993; Bardakci and Skibinski 1994), barbels mon (Verspoor and Hammar 1991) and genetic (Callejas and Ochando 1998) and a variety of other analysis is invaluable for determining hybrid fishes (Partis and Wells 1996). PCR amplification status; allozymes often make excellent diagnostic of 28S RNA from nDNA has been used to identify markers. F hybrids can be readily recognized as spiny lobster (Jasus edwardsii) phyllosome larvae 1 they are heterozygous for all nuclear loci discri- (Silberman and Walsh 1992). PCR-based analysis minating the two parental species. The female of DNA is also commonly used for identifying parent, and by elimination the male parent, of pathogenic and infectious organisms of fish. An known F hybrids can be identified using mtDNA example is mycobacterial identification using 1 markers. Backcross hybrids will be heterozygous RFLP analysis of 16S ribosomal RNA (Talaat et al. for some diagnostic loci but homozygous for 1997). A combination of mtDNA and nDNA others characterizing the species with which the analyses was used to track an individual hybrid F hybrid has crossed. Examples of the use of bio- blue/fin whale, Balaenoptera musculus ¥ B. 1 chemical genetics in hybridization studies are physalus, from the time of harpooning off Iceland given in Verspoor and Hammar (1991) and Ward in 1989 to sale in Japan as raw meat in 1993 and Grewe (1994); some recent examples of the (Cipriano and Palumbi 1999). Multilocus typing of joint use of nDNA and mtDNA markers include hypervariable microsatellite loci, although not the cyprinid genus Luxilus (Dowling et al. 1997), used in this study, lends itself particularly well to the salmonid genus Salvelinus (Wilson and identification and tracking of individuals. Bernatchez 1998), the percichthyid genus Genetic tests reveal that misidentification or Macquaria (Jerry et al. 1999) and the scorpaenid mislabelling of fish fillets is not uncommon. genus Sebastes (Seeb 1998). Huang et al. (1995) showed that 70% of a sample of claimed red snapper (Lutjanus campechanus) fillets in Florida were from congeneric species. Mislabelling of fish fillets in Australian outlets has 9.5 FISH POPULATION also been recorded (Partis and Wells 1996; R.D. GENETICS Ward, personal observation). Allozyme identifica- tion of material has been used in law courts in the Population genetics is a vast field. It includes USA (Harvey 1990) and Australia (C. Keenan, per- measurements of levels of variation within and sonal communication). J.E. Seeb et al. (1990) used between populations, the development of mathe- allozyme analysis to locate a king crab catch matical and simulation models of how genes 208 Chapter 9 mutate, differentiate and evolve, and studies of the lites; the monomorphic microsatellites, which forces changing gene frequencies, i.e. the relative admittedly are rather uncommon, tend to go roles of genetic drift and selection. Here I am unreported. The greater heterozygosity of mostly concerned with the assessment and inter- marine fish over freshwater fish may reflect their pretation of genetic variation within and between larger effective population sizes, which in turn fish populations. Nei (1987) and Hartl and Clark may reflect their larger and more connected (1997) provide good guides to the principles and environment. applications of population and evolutionary A comparison of the genetic make-up of genetics. different samples or subpopulations allows an Levels of genetic diversity vary among classes assessment of the degree of genetic connectivity of fish. Marine fish show higher levels of average between subpopulations (Fig. 9.2). A summary of allozyme heterozygosity (h) per locus than allozyme genetic diversity levels among samples anadromous fish (see Metcalfe et al., Chapter 8, or subpopulations of fish (Table 9.3) shows that, this volume), which in turn show higher heterozy- on average, about 5% of loci are expected to be gosity than freshwater fish (h = 0.059 vs. 0.052 vs. heterozygous in a sample (HS). The total heterozy- 0.046, respectively, Table 9.3, Ward et al. 1994). gosity (HT), that is, the expected heterozygosity de- The same pattern holds for microsatellite hetero- rived from allele frequencies averaged across zygosities (average marine h = 0.79, 12 species; populations, is a little higher, at about 6%. The anadromous h = 0.68, 7 species; freshwater h = proportion of genetic diversity attributable to

0.46, 13 species; see DeWoody and Avise 2000). differences among samples is the statistic FST (or Microsatellite heterozygosities are an order of GST), and can be estimated by (HT - HS)/HT (see magnitude greater than allozyme heterozygosi- Wright 1969 and Nei 1987 for further details). ties. This is largely because microsatellites are Where there is no differentiation among samples, intrinsically more variable than allozymes, with FST will be zero. Across all fish species, FST is about much higher mutation rates. In addition, the 0.13, indicating that about 13% of the observed allozyme datasets include monomorphic loci allele frequency variation arises from population whereas microsatellite studies are generally differentiation. biased towards the more polymorphic microsatel- There is a striking difference in the proportion

Table 9.3 Comparison of allozyme diversity levels among different classes of animals and between different groups of fish. Details are given of number of species, mean number of subpopulations sampled per species, mean number of loci sampled per species, and mean values of HT (total heterozygosity), HS (average sample heterozygosity) and FST (proportion of total genetic diversity attributable to sample differentiation). Standard errors are given. See text for further explanation. Fish data are from Ward et al. (1994); data for other vertebrates are updates of information in Ward et al. (1992). Requirements for inclusion are that 15 or more loci have been examined in 15 or more individuals per sample or subpopulation.

Class Number of species Number of subpopulations Number of loci HT HS FST

Fish 113 6.60 ± 0.46 28.08 ± 0.81 0.062 ± 0.004 0.053 ± 0.003 0.132 ± 0.016 Marine 57 6.40 ± 0.62 29.19 ± 1.22 0.064 ± 0.004 0.059 ± 0.004 0.062 ± 0.011 Anadromous 7 13.14 ± 3.12 27.43 ± 1.86 0.057 ± 0.007 0.052 ± 0.008 0.108 ± 0.044 Freshwater 49 5.90 ± 0.53 26.88 ± 1.18 0.062 ± 0.007 0.046 ± 0.005 0.222 ± 0.031 Amphibia 49 5.53 ± 0.57 22.59 ± 0.75 0.146 ± 0.012 0.097 ± 0.008 0.308 ± 0.033 Birds 28 4.75 ± 0.65 29.21 ± 1.34 0.059 ± 0.006 0.054 ± 0.006 0.078 ± 0.021 Mammals 83 4.82 ± 0.37 27.89 ± 0.85 0.077 ± 0.005 0.057 ± 0.004 0.207 ± 0.023 Reptiles 33 5.58 ± 0.73 23.73 ± 1.36 0.115 ± 0.014 0.086 ± 0.010 0.222 ± 0.036 Genetics 209

(a) undoubtedly reflect differences in levels of gene flow. Freshwater species occupy lakes or isolated No migration river catchments; gene flow rates between such populations are expected to be low and levels of differentiation are expected to be high. Marine fish occupy habitats with fewer barriers to dispersal, (b) and many have long-lived pelagic larvae that can drift widely in current systems. As a result, gene Limited migration flow levels among populations are expected to be high. The expected negative correlation between

dispersal ability and FST has been demonstrated for ten marine shorefishes and seven coral-reef fish (c) species by Waples (1987) and Doherty et al. (1995) respectively. However, not all marine larvae dis- perse over long distances and current eddies may Free migration assist larval retention; the recruitment of locally spawned larvae, even of coral-reef fish, can be sub- stantial (Jones et al. 1999; Swearer et al. 1999). This is of particular relevance to the value of marine Fig. 9.2 A simplified scheme of three types of migration protected areas as described by Polunin (Chapter pattern between two fished areas (denoted by shaded 14, Volume 2). boxes). In (a) interchange is effectively zero. An example The mean FST of marine fish is similar to that would be lake populations of freshwater fish; FST would of birds, while amphibia, which are generally re- be expected to be greater than 0.1. In (b) there is limited stricted to ponds and lakes for breeding, show even migration. An example would be fish populations occupying different tributaries of the same catchment, more population differentiation than freshwater fish (Table 9.3). Clearly dispersal ability does have or marine fish from widely separated areas; FST would be expected to approximate the 0.01–0.10 zone. In (c) there a major role in determining the degree of genetic is panmixia. An example would be fish from a single connectivity between populations. spawning site dispersing to different fishing grounds, The data in Table 9.3 are based on allozyme or a widely ranging non-homing migratory species; analysis; there are not, as yet, similar F sum- F would not be significantly different from zero. In ST ST maries using mtDNA or microsatellite data. practice, the size of FST would depend on the extent of contemporary and historical gene flow. Freshwater fish show generally higher levels of mtDNA haplotype divergence than marine species (Billington and Hebert 1991), but inter- population summaries are lacking. Carvalho and of genetic variation attributable to sample differ- Hauser (1998) list FST values from microsatellite ences between marine and freshwater species. In studies of five marine fish species; the mean is marine species, the mean FST is 0.062, much less 0.017 (range 0.035–0.006), less than the mean than the 0.222 of freshwater species (Table 9.3). In value of the marine fish allozyme data but similar fact, these mean values are inflated by the rela- to the median. FST values were tabled for only one tively high values of a few species: the median FST freshwater and two anadromous species. These ad- value for these 57 marine species is only 0.020, mittedly very limited data indicate that FST values one-seventh that of the 49 freshwater species at for microsatellites may well be rather similar to 0.144 (Waples 1998). Anadromous fish show inter- those from allozymes. This should not be surpris- mediate levels of differentiation. ing, as at equilibrium FST is approximately equal to These differences in levels of differentiation be- 1/(1 + 4Nem) (Wright 1969), where Ne is effective tween populations of marine and freshwater fish population size and m is migration rate. FST is then 210 Chapter 9 almost entirely dependent on the absolute from signal, a particular problem for marine numbers of migrants (Nem). However, if mutation species showing little differentiation, is to repli- rates (u) are high relative to gene flow rates, m may cate samples over time. be substituted by (m + u) (Wright 1969). In such a FST statistics are based on genotype and allele situation, increased population differentiation frequencies. With microsatellites, allele size infor- might be detected for loci with high mutation mation, determined as the number of repeat units, rates, such as microsatellites. Be that as it may, the is also available. Alleles with similar sizes at a particular advantages of microsatellites to popula- locus are considered to be genetically more similar tion analysis are probably in providing high levels than alleles with disparate sizes, and this size, or of polymorphism even in species with low levels of genetic relatedness, information along with fre- allozyme polymorphism, so providing enhanced quency information can be included in the analyti- power to detect differentiation, and in providing cal genetic distance framework. Such an analysis markers that are more likely to be selectively of population subdivision is an RST analysis neutral than allozymes. (Slatkin 1995; Goodman 1997). However, RST vari- The relationship between FST and numbers of ances are higher than FST variances (Slatkin 1995) migrants, Nem, is beguilingly simple. However, it and RST may be more sensitive to unequal sample is based on an island model of migration and only sizes (Ruzzante 1998). In addition, RST analyses holds under certain conditions. Some of these con- assume that microsatellite mutations occur in a ditions are that m has to be much less than 1 and stepwise size fashion, but the validity of this as- the same for each population, Ne has to be constant sumption is questionable. RST and FST analyses can both spatially and temporally, alleles must be yield different statistical results (e.g. O’Connell selectively neutral, and populations must be at et al. 1998; Shaw et al. 1999). Because of these equilibrium between migration and drift. Rarely and other issues, O’Connell and Wright (1997) if ever will these conditions be met. For example, recommended that analyses of microsatellite populations of fish are constantly expanding and variability in fishes be based on conventional FST contracting, and sometimes merge or separate. statistics. The time required to go halfway to equilibrium is The above discussion focuses on the use of 1 approximately 2m + –2Ne (Crow 1986). Therefore, FST methods for partitioning genetic variation. while small populations approach the new equilib- Recently, new approaches for studying migration rium quite quickly following a change in state, patterns have been proposed that do not assume large populations, which are typical of many symmetrical migration rates or equal population marine species, will take a long time. Cases of sizes. These are often based on maximum likeli- equilibrium may well be rare. Clearly, Nem esti- hood methods and coalescence theory (e.g. Neigel mates derived from FST estimates are at best ex- 1997; Beerli 1998; Luikart and England 1999) and pected to be crude approximations, and frequently are expected to provide improved estimates of may have little basis in reality. They become even population parameters. less meaningful when the large standard errors at- tached to F , which reflect sampling error from ST 9.5.1 Stock structure analysis numbers of individuals and numbers of loci, are translated into standard errors attached to Nem. In From a fisheries manager’s point of view, a stock such a situation, even when all the assumptions of is a group of fish that will react independently to the relationship hold, distinguishing a true Nem of exploitation from other such groups and which 10 (FST = 0.024) from a true Nem of 1000 (FST = consequently should be managed independently. 0.00024) can be problematic. These issues are dis- In practice, assessments of stock structure are dif- cussed further by Waples (1998) and Bossart and ficult and often contentious. Different types of in- Prowell (1998). Waples (1998) suggests that the formation may appear to delineate different stocks best single strategy for distinguishing FST noise and may contradict existing management bound- Genetics 211 aries. Stocks may expand or collapse, and the rivers to spawn. Determining the stock composi- dynamics of stock interactions are rarely if ever tion of these frequently fished mixed aggregations understood. However, managers ignore stock is an important management objective. Once structure issues at their peril and must consider genetic stock differences have been described, such information, even if incomplete, in their stock components in mixtures can be quantified stock assessments. using maximum likelihood mixed-stock analysis Perhaps the most commonly quoted biological (MSA; see Shaklee et al. 1999 for references). definition of a stock is that of Ihssen et al. (1981): ‘a Much of the chum salmon MSA is based on a stock is an intraspecific group of randomly mating database comprising approximately 20 allozyme individuals with temporal and spatial integrity’. loci typed in more than 250 collections from Japan, The implication is that there is restricted gene Russia, Alaska, British Columbia and Washington, flow with other such groups and that there is with sample sizes generally in the vicinity of 100 genetic differentiation between stocks. These per collection (Seeb and Crane 1999a). There is a issues are considered further in, for example, northern lineage with three genetic subgroups, Carvalho and Hauser (1994). However, marine fish Japan, Russia and northwestern Alaska, and a generally show less genetic population differentia- southern lineage with three subgroups, Alaskan tion than anadromous species, which in turn show Peninsula/Kodiak, southeastern Alaska/northern less differentiation than freshwater populations British Columbia/Prince William Sound and (Table 9.3) (Metcalfe et al., Chapter 8, this volume). southern British Columbia/Washington. A de- Consequently, determining the stock structure of tailed examination of western Alaska and Alaskan marine species is more challenging than for Peninsula populations using 40 allozyme loci re- anadromous or freshwater species. Findings of vealed an overall FST of 0.061 for this region alone no significant genetic differentiation between (Seeb and Crane 1999b). samples of commercial marine fish are quite There is also a less comprehensive mtDNA common (e.g. Ward and Elliott 2001). Such results database. This shows that a particular mtDNA are of little value to managers, being consistent haplotype is very common in Japan (~0.80), un- with models of population structure ranging from common in Russia and Alaska (frequencies of 0.13 complete mixing (panmixia) to 1% or less ex- and 0.07 respectively) and absent from British Co- change. Some examples of genetically based fish lumbia and Washington (Park et al. 1993). stock structure analysis follow. About 2000 chum salmon caught incidentally in Alaskan Peninsula sockeye salmon fisheries in 1993 and 1994 were genotyped for allozymes and Chum salmon (Oncorhynchus keta) mtDNA (Seeb and Crane 1999a). The subsequent Atlantic and Pacific salmonids have been targeted MSA estimated that most (~60%) were bound for for population/stock analysis, largely because of northwestern Alaska streams but stocks from their commercial importance and the threatened Japan (~15%) and from other Asian, Alaskan, nature of some stocks and also because of their British Columbian and Washington stocks were interesting anadromous life histories. Shaklee also harvested (Table 9.4). This information is et al. (1999) review how genetic data have been currently being used in conservation and manage- used by agencies in the Pacific northwest to ment decisions. manage four species of Pacific salmon; here one of At present there has been relatively little work these, the chum salmon, is considered. with microsatellites in chum salmon. However, Chum salmon have the most extensive natural eight populations (n > 50 per population) from the distribution of Pacific salmonids, ranging from Yukon River catchment were examined for 24 Korea and Japan north to Alaska and south to polymorphic nDNA loci (18 allozymes, one gene Oregon. They mature in the North Pacific, form- intron, five microsatellites) and for mtDNA ing mixed aggregations before returning to natal (Scribner et al. 1998). As expected, the different 212 Chapter 9

Table 9.4 Estimated contributions of Pacific Rim chum salmon from different regions to an Alaskan Peninsula fishery (Seeb and Crane 1999a). Standard deviations in parentheses. Sample sizes from 395 to 399 per collection. Fish were examined for 23 allozyme loci. The 1994 sample 21–25 June was also examined for mitochondrial DNA variation.

Region 1993 1994

13–20 June 22–29 June 17–20 June 21–25 June 26–30 June

Japan 0.17 (0.05) 0.15 (0.04) 0.13 (0.04) 0.09 (0.03) 0.16 (0.04) Russia 0.06 (0.07) 0.07 (0.07) 0.19 (0.07) 0.06 (0.09) 0.15 (0.07) NW Alaska (summer) 0.62 (0.10) 0.57 (0.10) 0.51 (0.09) 0.72 (0.10) 0.54 (0.08) Yukon (autumn) 0.00 (0.01) 0.01 (0.01) 0.00 (0.00) 0.00 (0.00) 0.02 (0.02) Alaskan Peninsula 0.04 (0.03) 0.13 (0.05) 0.07 (0.05) 0.10 (0.05) 0.05 (0.04) SE Alaska 0.06 (0.05) 0.05 (0.02) 0.03 (0.01) 0.02 (0.01) 0.04 (0.01) British Columbia 0.04 (0.02) 0.03 (0.01) 0.04 (0.02) 0.00 (0.01) 0.03 (0.02) Washington 0.01 (0.01) 0.00 (0.01) 0.02 (0.02) 0.00 (0.01) 0.00 (0.01) Unknown 0.003 0.000 0.003 0.000 0.005

markers showed different average heterozygosi- 1995), which suggested that gene flow was high ties (haplotype diversity for mtDNA): allozymes enough to prevent the formation of genetically dif- 0.250, nDNA 0.592, mtDNA 0.063. Estimates of ferentiated stocks below the scale of ocean basins.

FST were however highly concordant across mark- Conclusions from mtDNA studies have ranged er classes: allozymes 0.010 ± 0.003, nDNA 0.011 ± from a lack of differentiation across the Atlantic

0.004 and mtDNA 0.016. These FST estimates were (Smith et al. 1989) to heterogeneity and restricted significantly greater than zero, and were largely at- gene flow between east and west (Carr and tributable to differences between drainages in the Marshall 1991; Dahle 1991). Pogson et al. (1995) Yukon catchment. The addition of microsatellite examined anonymous nDNA markers by RFLP data to the existing allozyme and mtDNA data- analysis and found significant differentiation base will increase MSA power but will be costly. among Atlantic samples (FST = 0.069). They sug- gested that the allozyme homogeneity reflected balancing selection for expressed loci, while the Cod (Gadus morhua) RFLP heterogeneity reflected restricted gene flow The Atlantic cod is a demersal marine fish that in of neutral markers. Mork et al. (1985) had sug- 1997 was the ninth largest single-species fishery, gested effectively the reverse: the homogeneity of at 1.36 million tonnes (FAO 1999). Despite the most allozymes reflected extensive gene flow, the importance of the resource and detailed stock heterogeneity of a few allozymes reflected selec- assessments, a population collapse caused its tion. Galvin et al. (1995) examined a single commercial extinction off Newfoundland and minisatellite locus, finding significant allele

Labrador in 1992 (Hutchings and Myers 1994). frequency differences across the Atlantic (FST = Concern about the North Atlantic cod fisheries 0.03) and significant differences between New- has stimulated numerous genetic examinations foundland and Nova Scotia. The overall picture of population structure (see Shaklee and Bentzen then is of limited but significant genetic differenti- 1998). ation within the North Atlantic, with allozymes The largest allozyme survey (Mork et al. 1985) showing less differentiation than nDNA markers. found very limited differentiation between east There have also been detailed examinations of and west Atlantic cod (FST = 0.014; see Pogson et al. the stock structure of particular regions of the Genetics 213

Atlantic, in particular the Norwegian coast. Here northwestern samples: allele frequencies at only interest has focused on ‘arctic’ cod living in the one of six microsatellites were distinguishable Barents Sea in the northeastern Arctic and (Bentzen et al. 1996). It was suggested that conver- ‘coastal’ cod inhabiting Norwegian coastal and gent mutations at these highly mutable loci might fjord areas. These cod spawn in the same period have erased differentiation more readily detected and on the same grounds. They differ in allele fre- at less mutable markers such as allozymes and quencies for HbI (haemoglobin) (Frydenberg et al. mtDNA, and that therefore microsatellites might 1965; Dahle and Jørstad 1993), for an intron of be less suitable for broad-scale surveys than SypI (synaptophysin) (Fevolden and Pogson 1997) allozymes or mtDNA. and in four microsatellite loci (Dahle 1995). One One important point to be made from these allozyme locus, Ldh-3, showed differentiation, but various analyses is that relying on data from one up to 12 other allozymes (Mork et al. 1985; Mork marker or even one class of marker can lead to erro- and Giaever 1999) and ten cDNA clones (Fevolden neous or misleading conclusions. It is clearly desir- and Pogson 1995) did not; no mtDNA differentia- able to have data from as many markers as tion was detected (Arnason and Palsson 1996). The possible, yet there is then the possibility of differ-

SypI differentiation was remarkably high (FST = ent markers appearing to give conflicting results. 0.40), with evidence of selection at it or a linked Interpretation of likely population structures can region (Fevolden and Pogson 1997); interestingly, become difficult or controversial. More needs to be rather similar levels of SypI differentiation were known about the relative levels of selection and observed between two groups of Icelandic cod (FST mutation for different classes of marker, and this = 0.20; Jónsdóttir et al. 1999). Ldh-3 and HbI might information is rarely available. Markers that are also be subject to selection (Mork and Giaever neutral with respect to selection are desirable. 1999). It is not easy to reconcile these data into While microsatellites probably fit this require- a single satisfying and unambiguous picture of ment, they have some disadvantages such as some- population structure, and at present levels of gene times being too variable and/or difficult to score flow among populations in this region remain de- unambiguously; furthermore their underlying batable. For details of the physiological conse- mutation processes and consequent effects on pop- quences of the different haemoglobin genotypes, ulation differentiation need to be understood see Brix, Chapter 4, this volume. better. Another region to have attracted considerable attention is the northwestern Atlantic (Ruzzante Yellowfin tuna (Thunnus albacares) et al. 1999). Four to six microsatellite loci were examined in samples of larval cod, inshore and The yellowfin tuna is another valuable fishery, offshore cod, and cod from adjacent banks. ranking 11th in 1997 with 1.13 million tonnes Significant fine-scale differentiation has been de- (FAO 1999). The yellowfin tuna, like other tected, although FST levels were very low (gener- tunas, is a fast-swimming highly migratory pelagic ally, over all loci, <0.02, RST values tended to be a species, whose genetic population structure in the little higher). It was suggested that oceanographic major oceans has been quite extensively studied. structures, in the form of gyres and currents, and Some genetic diversity parameters are summa- spatiotemporal differences in spawning were bar- rized in Table 9.5. Portions of the data permitted riers to gene flow between neighbouring and even a hierarchical analysis of genetic variation, with contiguous populations. This evidence of small- the overall genetic diversity partitioned into scale stock structuring should be considered in within- and between-ocean components. The management plans aimed at rebuilding these within-ocean component almost entirely reflects depleted stocks. Counter-intuitively, the single re- a within-Pacific component, there being fewer gion examined from the eastern Atlantic (Barents samples from the Atlantic and Indian oceans.

Sea) showed little differentiation from these All FST estimates are low and many are not sta- 214 Chapter 9

Table 9.5 Heterozygosity and FST estimates for yellowfin tuna.

Oceans Marker Het No. of collections FST FSO FOT Reference

P, A MtDNA (12) 0.85 6 -0.021 -0.015 -0.006 Scoles and Graves (1993) P, I, A MtDNA (2) 0.68 9 0.010** 0.012* -0.002 Ward et al. (1997) P, I, A Allozyme (4) 0.36 8 0.025** 0.027** -0.002 Ward et al. (1997) P MtDNA (12) 0.85 5 -0.015 Scoles and Graves (1993) P MtDNA (2) 0.68 6 0.012* Ward et al. (1997), Grewe and Ward (unpublished data) P Allozyme (4) 0.36 6 0.027** Ward et al. (1997) P Microsatellites (6) 0.78 5–8 0.002* Grave and Ward (unpublished data) I MtDNA (2) 0.71 2 0.009 Ward et al. (1997)

P, Pacific Ocean; A, Atlantic Ocean; I, Indian Ocean. Marker: mtDNA (whole genome assayed by RFLP methods, with the given number of restriction enzymes), allozyme and microsatellites (number of polymorphic loci examined). Het, overall heterozygosity (genetic diversity for mtDNA). FST, proportion of total genetic variation attributable to sample differences; FSO, proportion of within-ocean genetic variation attributable to sample differences; FOT, proportion of total genetic variation attributable to ocean differences. The FST parameter for the single-ocean studies is analogous to the FSO parameter for the multiple-ocean studies. F estimates are from Hauser and Ward (1998) and used the AMOVA package (Excoffier et al. 1992). Significance of F parameters: *, P < 0.05; **, P < 0.01. tistically significant (Table 9.5). The maximum (Scoles and Graves 1993; Ward et al. 1997), al-

FST value for mtDNA was a barely significant lozyme (Ward et al. 1997) and microsatellite (Ward 0.012, and this related to differentiation within the and Grewe, unpublished data) surveys have shown Pacific Ocean; there was no detectable differentia- very little differentiation for any marker other tion between oceans. The maximum FST value than the allozyme Gpi-A. The FST value for this over the four polymorphic allozymes examined locus is about 0.10; other markers are 0.01 or less. was 0.027; again this was within the Pacific Ocean Perhaps the most likely interpretation of these and again there was no detectable differentiation data is that the Gpi-A differentiation within the between oceans. In fact, this allozyme differentia- Pacific reflects selection. The low but just statisti- tion reflects differentiation within the Pacific cally significant differentiation of mtDNA and Ocean of only one (Gpi-A) of four polymorphic microsatellites suggests that there might be some loci, with eastern samples being separable from weak genetic structuring of the Pacific Ocean western-central samples (Ward et al. 1997). So far yellowfin population but that, in general, gene microsatellite variation has only been examined flow is preventing any strong signals. in the Pacific Ocean, where a preliminary data ex- It should be noted that, contrary to yellowfin, amination shows a very low FST of about 0.002. there is striking mtDNA differentiation be- The finding of no differentiation for either tween Pacific and Atlantic populations of albacore mtDNA or allozymes between oceans suggests (Thunnus alalunga) and bigeye (Thunnus obesus) either that there is significant gene flow between tunas (Chow and Ushiama 1995; Alvarado Bremer the Atlantic and Indo-Pacific oceans, and yel- et al. 1998). Their interocean FST values are about lowfin tuna are indeed found around the Cape of 0.10 and 0.20 respectively (see Hauser and Ward Good Hope, or that the sampling regime has failed 1998). Extrapolation of population structure from to detect the existence of presumably low amounts one species to another, even when they have of interoceanic differentiation. similar ecologies, ranges and life-history attribut- Within the Pacific Ocean, extensive mtDNA es, is therefore fraught with danger (Graves 1998). Genetics 215

Furthermore, for the bigeye tuna, analyses of Some weak salmonid subpopulations in a mixed seven microsatellite loci and one nDNA locus (a ocean fishery could drop to small Ne values, where creatine kinase intron) have shown very little Ne is the biologically effective population size. For- Atlantic–Indo-Pacific differentiation (R.D. Ward, tunately, most wild fisheries are of marine species personal observation), although so far only about where populations are generally very large and 100 specimens have been examined. Yet again, where there is appreciable gene flow among popu- different marker classes appear to reveal different lations. In such a situation, fishing is unlikely to patterns of population heterogeneity. Consistent cause such a massive decrease in population size differences between nDNA and mtDNA patterns that an experimentally detectable amount of vari- might reflect the greater sensitivity of mtDNA ability will be lost by drift, at least given sample variants to bottleneck effects or sex differences in sizes of the order of 100. A fishery that did reach dispersal. low population levels would likely become uneco- nomic to exploit, although bycatch harvesting 9.5.2 Effects of fishing on could continue as other species become targeted. genetic diversity The major goal of fisheries management is to maintain populations at levels permitting sustain- Genetic diversity is a rich and irreplaceable part of able exploitation. However, it would be erroneous our heritage. The loss of genetic diversity is highly to consider marine species as ‘immune’ to genetic undesirable, whether the loss is in the form of depletion (see Ryman et al. 1995 for a full discus- species extinction or of intraspecific variability. sion), especially if the genetically effective popula- Can fishing itself lead to the loss of genetic varia- tion size of some species is much less than the tion? The answer must be yes, since fishing re- apparent size (Hedgecock 1994). duces population size from pre-fishing levels and There was some early evidence of a decrease some variation will therefore be lost by genetic in allozyme heterozygosity concomitant with drift (see also Reynolds et al., Chapter 15, Volume fishing pressure in the large deepwater orange 2). roughy (Hoplostethus atlanticus) fisheries off In fact, heterozygosity is lost relatively slowly. New Zealand. This was not attributed to a popula- The expected proportion of the original heterozy- tion bottleneck but to fishing removing larger and gosity remaining after a bottleneck of size N for putatively more heterozygous fish (Smith et al. 1 one generation is 1 - –2N. A single male and female 1991). However, there is no clear relationship be- will thus retain 75% of original heterozygosity, al- tween size and heterozygosity in orange roughy though the loss of heterozygosity compounds with (Ward and Elliott 1993), and studies over a longer successive generations of size N. A short-term time period found no correlation between fishing bottleneck will have much less drastic effects than pressure and genetic heterozygosity (Smith and a long-term bottleneck. Allele loss, however, will Benson 1997). There was no detectable change in be more striking. Regardless of how many alleles the allozyme heterozygosity of Hawaiian popula- existed at a locus in an ancestral population, a tions of the spiny lobster (Panulirus marginatus) single male and female cannot retain more than either before or after expansion of the fishery (L.W. four alleles. Rare alleles are likely to be rapidly lost Seeb et al. 1990). in a bottleneck. Loss of mtDNA diversity has been observed Loss of diversity is commonly observed in for the New Zealand endemic species Hector’s hatchery fish and shellfish; it is an expected dolphin (Cephalorhynchus hectori) (Pichler and consequence of genetic drift in small numbers of Baker 2000). A PCR-sequencing protocol was used broodstock. Significant losses arising from fishing to compare the mtDNA of museum and contem- pressures will be more noticeable in populations porary specimens. The North Island population that are already relatively small before fishing; over the last 20 years has declined from three freshwater fish could be particularly vulnerable. mtDNA lineages (diversity, h = 0.41) to one (h = 0), 216 Chapter 9 and the east coast South Island population from a single captive population appears free of foreign nine lineages (h = 0.65) to five (h = 0.35). Note that genes (Echelle and Echelle 1997). Introduced values of h can in principle range from 0, where all species may outcompete native species to the individuals are identical, to 1, where all indivi- point that inbreeding becomes a problem in the duals are different. It was suggested that this loss native species – an indirect genetic effect. of diversity resulted from a population decline Stock enhancement programmes do not intro- caused by dolphin entanglement in gill-nets, a loss duce exotic species but can nevertheless have exacerbated by very low gene-flow rates among unintended deleterious effects. Such programmes populations. The mtDNA FST of four populations may reintroduce strains from the target area or was high at 0.47, implying little population mix- introduce strains from another part of the species’ ing). Mitochondrial DNA diversity is especially range. These strains are very likely to hybridize sensitive to drift, as a single male and female mat- and introgress with native fish. Again, the fate of ing will transmit only one mtDNA haplotype, and introduced fish can be monitored if they carry the small fragmented populations of this species unique strain-specific markers. Allozyme studies make it particularly vulnerable to genetic loss. have confirmed introgression, for example in European brown trout (Salmo trutta) restocking 9.5.3 Genetic effects of stocking programmes. In Spain, local stock heterogeneity is being eroded by introductions of hatchery-reared There are two major classes of stocking pro- fish (Machordom et al. 1999), and in Switzerland grammes for fish: introduced trout of Atlantic basin origin are re- 1 those where exotic species are introduced to placing native Adriatic strains S. trutta fario and S. areas where they have never naturally existed; trutta marmoratus (Largiader and Scholl 1995). 2 those designed to enhance pre-existing fisheries. Clearly, where possible, enhancement should be The latter may be enacted to replace a locally of local rather than exotic strains. Even then, the extinct stock, to rebuild a collapsed stock or to introduced strains are likely to be of hatchery augment a natural population for recreational fish- origin and therefore will have undergone some ing. While stocking programmes may have major change in genetic make-up with respect to ances- benefits, they may also have very significant detri- tral populations. mental outcomes (see Cowx, Chapter 17, Volume One undesirable consequence of such 2). Introduction of exotic species can lead to the hybridization is that locally coadapted gene com- elimination of native species, hybridization with plexes, which will be a particular genetic make- native species, habitat alteration and the spread up favoured by selection in that location, can of disease. Stock enhancement programmes can break down, reducing fitness. Such outbreeding introduce disease and can deleteriously affect the depression has been recorded for pink salmon gene pool of native fish. Cowx (1998) provides a (Oncorhynchus gorbuscha) (Gharrett and Smoker recent collection of papers on fish stocking. 1991). Hatchery-bred rainbow trout, marked by a Introductions of exotic species to habitats rare allozyme variant, had a reproductive success occupied by a closely related species can lead to only 28% that of wild trout, but so outnumbered hybridization and introgression, problems readily native fish that 62% of smolts were offspring of monitored by species-diagnostic loci. In Arizona, naturally spawning hatchery trout (Chilcote et al. the gene pool of the Apache trout (Oncorhynchus 1986). There is ‘compelling evidence’ that apache) has been threatened by hybridizations hatchery production of at least three species of with introduced non-native cutthroat (O. clarki) Pacific salmonids lowers fitness, and that when and rainbow (O. mykiss) trout (Carmichael et al. such fish hybridize with wild fish, the productivity 1993). All wild populations of the leon springs pup- and viability of the naturally spawning popula- fish (Cyprinodon bovinus) have been introgressed tions declines substantially (Reisenbichler and by introduced sheepshead minnow (C. variegatus); Rubin 1999). Genetics 217

There may also be indirect genetic effects of within a species or how this diversity is parti- enhancement programmes. One example con- tioned among populations. Without this informa- cerns introductions of Atlantic salmon (Salmo tion, it is not possible to assess the effectiveness or salar) genetically resistant to but carrying an ec- appropriateness of conservation measures. Some toparasite; the recipient population was suscepti- quantification of the extent of genetic population ble, causing heavy mortality of native fish (Bakke structuring of threatened or vulnerable species is 1991). therefore desirable. However, if this structuring is A particular threat to the genetic integrity and extensive, as is frequently the case for freshwater viability of wild fish populations may come from and anadromous species but less so for marine the deliberate or accidental release of genetically species (Ward et al. 1994), it may not be feasible to engineered or transgenic fish (Reichhardt 2000). enact conservation measures to protect all compo- Such fish are likely to be engineered for higher nents. Pacific salmonids, for example, are sepa- growth rate by elevating growth hormone (GH) rated into numerous genetically distinct levels. The effects of such fish on native popula- populations through their strong homing behav- tions remain unknown. They are often assumed iour. Allendorf et al. (1997) consider that more to have reduced viability, but GH-enhanced than 300 native stocks of these salmonids are at transgenic coho salmon (Oncorhynchus kisutch) risk in the Pacific northwest and suggest how they showed a marked competitive feeding advantage may be prioritized for conservation actions using over non-transgenic fish (Devlin et al. 1999). Thus, ranks based on the genetic and ecological conse- depending of course on other fitness-related traits, quences of extinction. Such actions might include escaped GH transgenic fish may be able to compete habitat protection, reduction in harvesting rate, successfully in the wild with native non- stock enhancement using strains from the stock to transgenic fish. Even if the transgene does lower be protected or, if that is not possible, the most viability, modelling studies (Muir and Howard closely related population, or translocation to a 1999) have shown that it can still spread if it simul- new habitat. Some genetic effects associated with taneously increases male mating success, for ex- the last two issues have already been considered. ample by yielding faster-growing larger males. Another concern is the need to maintain genetic This can, in principle, lead to local extinction. It diversity levels as high as possible in any translo- seems that in order to minimize the genetic effects cated population. The minimum effective popula- of escapees on local populations, commercial pro- tion size of a hatchery or transplanted stock should ducers of transgenic fish will have to guarantee be, whenever possible, around 50–100 animals. A that the transgenic product is sterile. This might high population size and diversity permits some be achievable through the production of triploid measure of adaptive evolution and reduces the fish or by inclusion of a transgene for repressible deleterious effects of inbreeding. sterility (Grewe et al. submitted). Containment fa- Ryman et al. (1995) review the many threats to cilities would still be advisable; the release of large genetic diversity in fish and suggest three respons- numbers of voraciously feeding, rapidly growing, es to these threats: (i) to ensure that those involved GH-enhanced fish could seriously disrupt the in operations such as wild harvesting, aquaculture functioning of natural ecosystems, even if the fish and ocean ranching are aware of possible incom- were sterile. patibilities with conservation objectives; (ii) to require a risk assessment before any potentially 9.5.4 Conservation genetics threatening activity commences; and (iii) to accept that whoever proposes a threatening activity is The desirability of retaining or conserving genetic responsible for providing the burden of proof diversity is now widely accepted. Once lost, diver- of the likely impact of that action. Further aspects sity cannot be restored. Unfortunately, we often of conservation are discussed by Reynolds et al. have little knowledge of the extent of diversity (Chapter 15, Volume 2). 218 Chapter 9

et al. 1998) and inferred in Acipenser and Rhodeus 9.6 GENETICS OF SEX by ploidy and gynogenesis manipulations DETERMINATION IN FISH (Kawamura 1998; Van Eenennaam et al. 1999). The lake trout (Salvelinus namaycush) is one of the few Reproductive mechanisms in fish are highly salmonids with morphologically differentiated variable (reviewed by Devlin and Nagahama 2002). sex chromosomes, and the sex-determining region While most reproduce sexually, some are asexual. appears to lie on the short arm of the Y chromo- Some of the asexual species consist of all-female some (Reed et al. 1995). Given the difficulty of dis- gynogens, whereby matings with a male of a tinguishing sex chromosomes in most salmonids, related species are necessary to initiate embryoge- and the importance of sex manipulations in nesis but the male does not contribute genetic ma- aquaculture production, there have been consider- terial. Examples are the gynogenetic Poecilia able efforts put into isolating salmonid Y-specific formosa (sailfin molly), whose embryogenesis is probes. These are now available for five different initiated by sperm from Poecilia latipinna in Oncorhynchus species (Devlin et al. 1994; mixed-species shoals (Schlupp and Ryan 1996), Donaldson and Devlin 1996; Nakayama et al. and gynogenetic minnows of the Phoxinus eos- 1999), although not yet for Atlantic salmon. neogaeus complex. In the latter, examinations of Environmental sex determination, principally DNA variability have revealed only one clone or through temperature, has been suggested for a genotype in more than 400 gynogenetic fish sam- range of fish from lampreys (Beamish 1993) to pled (Elder and Schlosser 1995). Another type of eels (Krueger and Oliveira 1999) and atherinids asexual reproduction is hybridogenesis. In Poecil- (Strussmann et al. 1997). In the cichlid genus iopsis, the all-female hybridogenetic P. monacha- Apistogramma, pH as well as temperature is a lucida biotype arose as an interspecific hybrid of P. significant determinant of sex in some species monacha and P. lucida. Subsequent maintenance (Romer and Beisenherz 1996). There is evidence of this biotype relies on matings with males of P. that environmental factors often overlay or modify lucida, but the lucida genome is discarded during the outcomes of a genetic switch (e.g. Conover oogenesis (Schultz 1969; Vrijenhoek 1994). et al. 1992; Strussmann et al. 1997; Abucay et al. Among sexual fish species, there is a wide 1999). Finally, some fish are protandrous hermaph- range of sex-determining systems, including both rodites, with sex changing from male to female genetic and environmental mechanisms. Genetic with age (e.g. sea bass or barramundi, Lates calcar- mechanisms in fish are varied, with Tave (1993) ifer; Guiguen et al. 1993), while some are protogy- listing nine types. Fifteen fish species had the nous, with sex changing from female to male XX/XY chromosome type with homogametic (e.g. bluehead wrasse, Thalassoma bifasciatum; females and heterogametic males; five had the Kramer and Imbriano 1997). ZZ/ZW type with homogametic males and het- erogametic females; eight had other sex chro- mosome systems; and two had autosomal sex 9.7 CONCLUSIONS determination. The XX/XY system is found in many aquacultured species, including channel It should be clear from the discussion in this chap- catfish, salmonids and carp. Often the sex chromo- ter that genetic analyses have a great deal to offer in somes are not morphologically differentiated and the study of fish populations. The range of tools is cannot be identified by chromosome staining tech- ever-expanding as new ways to harness the power niques; in such cases the system has to be inferred of PCR analysis are brought on-line. These dif- from sex-reversal or other genome manipula- ferent tools are suited to different problems, and I tions. For example, a ZZ/ZW mechanism has been have tried to outline both their benefits and limita- shown in Leporinus and Clarias by direct chromo- tions. The focus has been on the uses of genetics in some staining (Pandey and Lakra 1997; Molina fish identification and population structure analy- Genetics 219 sis, but molecular genetic studies have also made numerous contributions to other areas such as sys- ACKNOWLEDGEMENTS tematics and aquaculture. I thank Sharon Appleyard, Nick Elliott and Dan Genetic identification of unknown or disputed McGoldrick for comments on this chapter. samples is usually unambiguous and simple; protein or mtDNA-based tests are very suitable. The analysis of population structure is far more demanding. Interpretation of frequently large and REFERENCES complex datasets can be difficult, and estimates of gene-flow levels are at best likely to be only Abucay, J.S., Mair, G.C., Skibinski, D.O.F. and Beardmore, J.A. (1999) Environmental sex determina- crude approximations to true values. Increasing tion: the effect of temperature and salinity on sex ratio the number of markers in a study will increase re- in Oreochromis niloticus L. Aquaculture 173, solving power, but the attributes of the various 219–34. markers need to be carefully considered in reach- Allendorf, F.W., Bayles, D., Bottom, D.L. et al. (1997) ing any conclusion. These include mode of inheri- Prioritizing Pacific salmon stocks for conservation. tance, mutation rate, degree of selective neutrality Conservation Biology 11, 140–52. Alvarado Bremer, J.R., Stequert, B., Robertson, N.W. and likelihood of scoring errors. Ideally, selec- and Ely, B. (1998) Genetic evidence for inter-oceanic tively neutral nDNA and mtDNA markers would subdivision of bigeye tuna (Thunnus obesus Lowe) be jointly deployed, in sample sizes of 100 or more. populations. Marine Biology 132, 547–57. Genetic analysis is of course not the only option Arnason, E. and Palsson, S. (1996) Mitochondrial for examining fish population structure. The cytochrome b DNA sequence variation of Atlantic cod many other tools that can be applied include Gadus morhua, from Norway. Molecular Ecology 5, examination of life-history traits and morpho- 715–24. Bakke, T.A. (1991) A review of the inter- and intraspecific metric and meristic variation, tagging returns variability in salmonid hosts to laboratory infections (Metcalfe et al., Chapter 8, this volume), parasite with Gyrodactylus salaris Malmberg. Aquaculture loads and microchemical variation of hard parts. 98, 303–10. The various genetic and non-genetic tools all have Bardakci, F. and Skibinski, D.O.F. (1994) Application of their attendant advantages and disadvantages but the RAPD technique in Tilapia fish: species and sub- all provide valuable and complementary informa- species identification. Heredity 73, 117–23. tion and every type of available information Beamish, F.W. (1993) Environmental sex determination in southern brook lamprey, Ichthyomyzon gagei. should be considered when drawing up models of Canadian Journal of Fisheries and Aquatic Sciences population structure. Physical data, such as ba- 50, 1299–307. thymetry and direction and strength of eddies and Beerli, P. (1998) Estimation of migration rates and popula- currents, should also be factored into likely popu- tion sizes in geographically structured populations. In: lation models. Such a coordinated multidiscipli- G.R. Carvalho (ed.) Advances in Molecular Ecology. nary approach is much to be encouraged, albeit Amsterdam: IOS Press, pp. 39–53. Begg, G.A., Friedland, K.D. and Pearce, J.B. (1999) Stock challenging, but may be too resource-demanding identification and its role in stock assessment and for studies of any other than the major fisheries. fisheries management: an overview. Fisheries Re- Finally, genetic diversity once lost from a search 43, 1–8. species can never be regained. Careful husbandry Bentzen, P., Taggart, C.T., Ruzzante, D.E. and Cook, D. of our genetic resources requires the continued (1996) Microsatellite polymorphism and the popula- deployment of genetic analyses for cataloguing, tion structure of cod (Gadus morhua) in the North monitoring and conservation. West Atlantic. Canadian Journal of Fisheries and Aquatic Sciences 53, 2706–21. Beverton, R.J.H. (1990) Small marine pelagic fish and the threat of fishing: are they endangered? Journal of Fish Biology 37 (Suppl. A), 5–16. Billington, N. and Hebert, P.D.N. (1991) Mitochondrial 220 Chapter 9

DNA diversity in fishes and its implications for intro- Conover, D.O., VanVoorhees, D.A. and Ehtisham, A. ductions. Canadian Journal of Fisheries and Aquatic (1992) Sex-ratio selection and the evolution of environ- Sciences 48 (Suppl. 1), 80–94. mental sex determination in laboratory populations of Bossart, J.L. and Prowell, D.P. (1998) Genetic estimates Menidia menidia. Evolution 46, 1722–30. of population structure and gene flow: limitations, Cowx, I.G. (ed.) (1998) Stocking and Introduction of Fish. lessons and new directions. Trends in Ecology and Oxford: Fishing News Books. Evolution 13, 202–6. Crow, J.F. (1986) Basic Concepts in Population, Caetano-Anollés, G. and Gresshoff, P.M. (eds) (1997) Quantitative, and Evolutionary Genetics. New York: DNA Markers: Protocols, Applications and Freeman. Overviews. New York: John Wiley & Sons. Dahle, G. (1991) Cod, Gadus morhua L., populations Callejas, C. and Ochando, M.D. (1998) Identification of identified by mitochondrial DNA. Journal of Fish Spanish barbel species using the RAPD technique. Biology 38, 295–303. Journal of Fish Biology 53, 208–15. Dahle, G. (1995) Genetic structure of the North-east At- Carmichael, G.J., Hanson, J.N., Schmidt, M.E. and lantic cod (Gadus morhua L.): an appraisal of different Morizot, D.C. (1993) Introgression among apache, cut- molecular techniques. PhD thesis, University of throat, and rainbow trout in Arizona. Transactions of Bergen, Norway. the American Fisheries Society 121, 26–35. Dahle, G. and Jørstad, K. (1993) Haemoglobin variation in Carr, S.M. and Marshall, H.D. (1991) A direct approach to cod: a reliable marker for Arctic cod (Gadus morhua the measurement of genetic variation in fish popula- L.). Fisheries Research 16, 301–11. tions: applications of the polymerase chain reaction to de Ligny, W. (1969) Serological and biochemical studies studies of Atlantic cod, Gadus morhua. Journal of Fish in fish populations. Oceanography and Marine Biology 39 (Suppl. A), 101–7. Biology Annual Review 7, 411–513. Carvalho, G.R. and Hauser, L. (1994) Molecular genetics Deriso R. and Quinn, T. (co-chairs) (1998) Improving and the stock concept in fisheries. Reviews in Fish Fish Stock Assessments. Washington, DC: National Biology and Fisheries 4, 326–50. Academy Press. Carvalho, G.R. and Hauser, L. (1998) Advances in DeSalle, R. and Birstein, V.J. (1996) PCR identification of the molecular analysis of fish population structure. black caviar. Nature 381, 197–8. Italian Journal of Zoology 65 (Suppl.), 21–33. Devlin, R.H., McNeil, B.K., Solar, I.I. and Donaldson, Cespedes, A., Garcia, T., Carrera, E. et al. (1998) Poly- E.M. (1994) A rapid PCR-based test for Y-chromosomal merase chain reaction restriction fragment length DNA allows simple production of all-female strains of polymorphism analysis of a short fragment of the cy- chinook salmon. Aquaculture 128, 211–20. tochrome b gene for identification of flatfish species. Devlin, R.H., Johnsson, J.I., Smailus, D.E., Biagi, C.A., Journal of Food Protection 61, 1684–5. Jonsson, E. and Bjornsson, B.T. (1999) Increased ability Chen, I.C., Chapman, F.A., Wei, C.I. and O’Keefe S.F. to compete for food by growth hormone-transgenic (1996) Preliminary studies on SDS-PAGE and isoelec- coho salmon Oncorhynchus kisutch (Walbaum). tric focusing identification of sturgeon sources of Aquaculture Research 30, 479–82. caviar. Journal of Food Science 61, 533. Devlin, R.H. and Nagahama, Y. (2002) Sex determination Chilcote, M.W., Leider, S.A. and Loch, J.J. (1986) and sex differentiation in fish: an overview of genetic, Differential reproductive success of hatchery and wild physiological, and environmental influences. Aqua- summer-run steelhead under natural conditions. culture 208, 191–364. Transactions of the American Fisheries Society 115, DeWoody, J.A. and Avise, J.C. (2000) Microsatellite varia- 726–35. tion in marine, freshwater and anadromous fishes Chow, S. (1994) Identification of billfish species using compared with other animals. Journal of Fish Biology mitochondrial cytochrome b gene fragment amplified 56, 461–73. by polymerase chain reaction. In: International Dinesh, K.R., Lim, T.M., Chua, K.L., Chan, W.K. and Commission for the Conservation of Atlantic Tunas. Phang, V.P.E. (1993) RAPD analysis: an efficient Report of the 2nd ICCAT Billfish Workshop. Madrid: method of DNA fingerprinting in fishes. Zoological International Commission for the Conservation of Science 10, 849–54. Atlantic Tunas, pp. 549–56. Doherty, P.J., Planes, S. and Mather, P. (1995) Gene flow Chow, S. and Ushiama, H. (1995) Global population and larval duration in seven species of fish from the structure of albacore (Thunnus alalunga) inferred by Great Barrier Reef. Ecology 76, 2373–91. RFLP analysis of the mitochondrial ATPase gene. Donaldson, E.M. and Devlin, R.H. (1996) Uses of biotech- Marine Biology 123, 39–45. nology to enhance production. In: W. Pennell and B.A. Cipriano, F. and Palumbi, S.R. (1999) Genetic tracking of Barton (eds) Principles of Salmonid Culture. Amster- a protected whale. Nature 397, 307–8. dam: Elsevier, pp. 969–1020. Genetics 221

Dowling, T.E., Broughton, R.E. and DeMarais, B.D. tion structures of cosmopolitan marine fishes. Journal (1997) Significant role for historical effects in the of Heredity 89, 427–37. evolution of reproductive isolation: evidence from Grewe, P.M., Patil, J.G., McGoldrick, D.J. et al. (2000) patterns of introgression between the cyprinid fishes, Preventing genetic pollution and the establishment of Luxilus cornutus and Luxilus chrysocephalus. Evolu- feral populations: a molecular solution. Reviews in tion 51, 1574–83. Fish Biology and Fisheries, submitted. Echelle, A.A. and Echelle, A.F. (1997) Genetic introgres- Guiguen, Y., Jalabert, B., Thouard, E. and Fostier, A. sion of endemic taxa by non-natives: a case study with (1993) Changes in plasma and gonadal steroid hor- leon springs pupfish and sheepshead minnow. Conser- mones in relation to the reproductive cycle and the sex vation Biology 11, 153–61. inversion process in the protandrous sea-bass, Lates Elder, J.F. and Schlosser, I.J. (1995) Extreme clonal unifor- calcarifer. General and Comparative Endocrinology mity of Phoxinus eos-neogaeus gynogens (Pisces, 92, 327–38. Cyprinidae) among variable habitats in northern Harris, H. (1966) Enzyme polymorphisms in man. Pro- Minnesota beaver ponds. Proceedings of the National ceedings of the Royal Society of London Series B 164, Academy of Sciences USA 92, 5001–5. 298–310. Etienne, M., Jerome, M., Fleurence, J. et al. (2000) Identi- Hartl, D.L. and Clark, A.G. (1997) Principles of fication of fish species after cooking by SDS-PAGE and Population Genetics, 3rd edn. Sunderland, MA: urea IEF: a collaborative study. Journal of Agricultural Sinauer Associates. and Food Chemistry 48, 2653–8. Harvey, W.D. (1990) Electrophoretic techniques in foren- Evans, B.S., White, R.W.G. and Ward, R.D. (1998) Genetic sics and law enforcement. In: D.H. Whitmore (ed.) identification of asteroid larvae from Tasmania, Aus- Electrophoretic and Isoelectric Focusing Techniques tralia, by PCR-RFLP. Molecular Ecology 7, 1077–82. in Fisheries Management. Boca Raton, FL: CRC Press, Excoffier, L., Smouse, P. and Quattro, J. (1992) Analysis of pp. 313–21. molecular variance inferred from metric distances Hauser, L. and Ward, R.D. (1998) Population identifica- among DNA haplotypes: application to human mito- tion in pelagic fish: the limits of molecular markers. In: chondrial DNA restriction data. Genetics 131, 479–91. G.R. Carvalho (ed.) Advances in Molecular Ecology. FAO (1999) FAO Yearbook 1997. Fishery Statistics: Amsterdam: IOS Press, pp. 191–224. Catches and Landings, Vol. 82. Rome: FAO. Hedgecock, D. (1994) Does variance in reproductive suc- Fevolden, S.E. and Pogson, G.H. (1995) Differences in cess limit effective population sizes of marine organ- nuclear DNA RFLPs between the Norwegian coastal isms? In: A.R. Beaumont (ed.) Genetics and Evolution and the North-east arctic populations of Atlantic cod. of Marine Organisms. London: Chapman & Hall, pp. In: H.R. Skjoldal, C. Hopkins, K.E. Erikstad and H.P. 122–34. Leinås (eds) Ecology of Fjords and Coastal Waters. Hillis, D.M., Moritz, C. and Mable, B.K. (eds) (1996) Mol- Amsterdam: Elsevier, pp. 403–15. ecular Systematics, 2nd edn. Sunderland, MA: Sinauer Fevolden, S.E. and Pogson, G.H. (1997) Genetic Associates. divergence at the synaptophysin (Syp I) locus among Hsieh, Y.H.P., Chen, F.C. and Nur, M. (1997) An hypothe- Norwegian coastal and north-east Arctic populations sis paper: rapid species identification of cooked red of Atlantic cod. Journal of Fish Biology 51, 895–908. snapper using isoelectric focusing. Journal of Food Frydenberg, O., Moller, D., Naevdal, G. and Sick, K. Science 62, 15–19. (1965) Haemoglobin polymorphism in Norwegian cod Huang, T.S., Marshall, M.R. and Wei, C.I. (1995) Identifi- populations. Hereditas 53, 257–71. cation of red snapper (Lutjanus campechanus) using Galvin, P., Sadusky, T., McGregor, D. and Cross, T. (1995) electrophoretic techniques. Journal of Food Science Population genetics of Atlantic cod using amplified 60, 279–83. single locus minisatellite VNTR analysis. Journal of Hutchings, J.A. and Myers, R.A. (1994) What can be Fish Biology 47 (Suppl. A), 200–8. learned from the collapse of a renewable resource: Gharrett, A.J. and Smoker, W.W. (1991) Two generations Atlantic cod, Gadus morhua, of Newfoundland and of hybrids between even- and odd-year pink salmon Labrador. Canadian Journal of Fisheries and Aquatic (Oncorhynchus gorbuscha): a test for outbreeding de- Sciences 51, 2126–46. pression. Canadian Journal of Fisheries and Aquatic Ihssen, P.E., Booke, H.E., Casselman, J.M., McGlade, Sciences 48, 1744–9. J.M., Payne, N.R. and Utter, F.M. (1981) Stock identifi- Goodman, S.J. (1997) Rst Calc: a collection of computer cation: materials and methods. Canadian Journal of programs for calculating estimates of genetic differen- Fisheries and Aquatic Sciences 38, 1838–55. tiation from microsatellite data and determining their Innes, B.H., Grewe, P.M. and Ward, R.D. (1998) significance. Molecular Ecology 6, 881–5. PCR-based genetic identification of marlin and other Graves, J.E. (1998) Molecular insights into the popula- billfish. Marine and Freshwater Research 49, 383–8. 222 Chapter 9

Jerry, D.R., Raadil, T.A., Cairns, S.C. and Baverstock, P.R. Mork, J. and Giaever, M. (1999) Genetic structure of cod (1999) Evidence for natural interspecific hybridization along the coast of Norway: results from isozyme between the Australian bass (Macquaria novemac- studies. Sarsia 84, 157–68. uleata) and estuary perch (M. colonorum). Marine and Mork, J.P., Ryman, N., Ståhl, G., Utter, F. and Sundnes, G. Freshwater Research 50, 661–6. (1985) Genetic variation in Atlantic cod (Gadus Jones, G.P., Milicich, M.J., Emslie, M.J. and Lunow, C. morhua) throughout its range. Canadian Journal of (1999) Self-recruitment in a coral reef fish population. Fisheries and Aquatic Sciences 42, 1580–7. Nature 402, 802–4. Mueller, U.G. and Wolfenbarger, L.L. (1999) AFLP Jónsdóttir, Ó.D.B., Imsland, A.K., Daníelsdóttir, A.K., genotyping and fingerprinting. Trends in Ecology and Thorsteinsson, V. and Naevdal, G. (1999) Genetic Evolution 14, 389–94. differentiation among Atlantic cod in south and Muir, W.M. and Howard, R.D. (1999) Possible ecological south-east Icelandic waters: synaptophysin (Syp I) and risks of transgenic organism release when transgenes haemoglobin (Hb I) variation. Journal of Fish Biology affect mating success: sexual selection and the Trojan 54, 1259–74. gene hypothesis. Proceedings of the National Kawamura, K. (1998) Sex determination system of the Academy of Sciences USA 96, 13853–6. rosy bitterling, Rhodeus ocellatus ocellatus. Environ- Nakayama, I., Biagi, C.A., Koide, N. and Devlin, R.H. mental Biology of Fishes 52, 251–60. (1999) Identification of a sex-linked GH pseudogene in Kramer, C.R. and Imbriano, M.A. (1997) Neuropeptide one of two species of Japanese salmon (Oncorhynchus (NPY) induces gonad reversal in the protogynous masou and O. rhodurus). Aquaculture 173, 65–72. bluehead wrasse, Thalassoma bifasciatum (Teleostei: Nei, M. (1987) Molecular Evolutionary Genetics. New Labridae). Journal of Experimental Zoology 279, York: Columbia University Press. 133–44. Neigel, J.E. (1997) A comparison of alternative strategies Krueger, W.H. and Oliveira, K. (1999) Evidence for for estimating gene flow from genetic markers. Annual environmental sex determination in the American eel, Review of Ecology and Systematics 28, 105–28. Anguilla rostrata. Environmental Biology of Fishes Nielsen, E.E., Hansen, M.M. and Loeschcke, V. (1999) 55, 381–9. Analysis of DNA from old scale samples: technical as- Largiader, C.R. and Scholl, A. (1995) Effects of stocking pects, applications and perspectives for conservation. on the genetic diversity of brown trout populations of Hereditas 130, 265–76. the Adriatic and Danubian drainages in Switzerland. O’Connell, M. and Wright, J.M. (1997) Microsatellite Journal of Fish Biology 47 (Suppl. A), 209–25. DNA in fishes. Reviews in Fish Biology and Fisheries Lewontin, R.C. (1974) The Genetic Basis of Evolutionary 7, 331–63. Change. New York: Columbia University Press. O’Connell, M., Dillon, M.C., Wright, J.M., Bentzen, P., Lewontin, R.C. and Hubby, J.L. (1966) A molecular Merkouri, S. and Seeb J. (1998) Genetic structuring approach to the study of genetic variation in natural among Alaskan Pacific herring (Clupea pallasi) popu- populations. II. Amount of variation and degree of lation identified using microsatellite variability. heterozygosity in natural populations of Drosophila Journal of Fish Biology 53, 150–63. pseudoobscura. Genetics 54, 595–609. Pandey, N. and Lakra, W.S. (1997) Evidence of female het- Luikart, G. and England, P.R. (1999) Statistical analysis erogamety, B-chromosome and natural tetraploidy in of microsatellite DNA data. Trends in Ecology and the Asian catfish, Clarias batrachus, used in aquacul- Evolution 14, 252–6. ture. Aquaculture 149, 31–7. Machordom, A., Garcia-Marin, J.L., Sanz, N., Almod- Park, L.K., Brainard, M.A., Dightman, D.A. and Winans, ovar, A. and Pla, C. (1999) Allozyme diversity in brown G.A. (1993) Low levels of intraspecific variation in the trout (Salmo trutta) from central Spain: genetic conse- mitochondrial DNA of chum salmon (Oncorhynchus quences of restocking. Freshwater Biology 41, 707–17. keta). Molecular Marine Biology and Biotechnology 2, Manchenko, G.P. (1994) Handbook of Detection of 362–70. Enzymes on Electrophoretic Gels. Boca Raton, FL: Partis, L. and Wells, R.J. (1996) Identification of fish CRC Press. species using random amplified polymorphic DNA McQuinn, I.H. (1997) Metapopulations and the Atlantic (RAPD). Molecular and Cellular Probes 10, 435–41. herring. Reviews in Fish Biology and Fisheries 7, Pichler, F.B. and Baker, C.S. (2000) Loss of genetic 297–329. diversity in the endemic Hector’s dolphin due to Molina, W.F., Schmid, M. and Galetti, P.M. (1998) Hete- fisheries-related mortality. Proceedings of the Royal rochromatin and sex chromosomes in the Neotropical Society of London Series B 267, 97–102. fish genus Leporinus (Characiformes, Anosromidae). Pogson, G.H., Mesa, K.A. and Boutilier, R.G. (1995) Cytobios 94, 141–9. Genetic population structure and gene flow in the Genetics 223

Atlantic cod Gadus morhua: a comparison of allozyme Schultz, R.J. (1969) Hybridization, unisexuality and and nuclear RFLP loci. Genetics 139, 375–85. polyploidy in the teleost Poeciliopsis (Poeciliidae) and Quattro, J.M. and Jones, W.J. (1999) Amplification other vertebrates. American Naturalist 103, 605–19. primers that target locus-specific introns in Scoles, D.R. and Graves, J.E. (1993) Genetic analysis of actinopterygian fishes. Copeia 1999, 191–6. the population structure of yellowfin tuna, Thunnus Reed, K.M., Bohlander, S.K. and Phillips, R.B. (1995) Mi- albacares, from the Pacific Ocean. Fishery Bulletin 91, crodissection of the Y-chromosome and fluorescence 690–8. in-situ hybridisation analysis of the sex chromosomes Scribner, K.T., Crane, P.A., Spearman, W.J. and Seeb, L.W. of lake trout, Salvelinus namaycush. Chromosome (1998) DNA and allozyme markers provide concordant Research 3, 221–6. estimates of population differentiation: analyses of Rehbein, H., Mackie, I.M., Pryde, S. et al. (1999) Fish U.S. and Canadian populations of Yukon River fall-run species identification in canned tuna by PCR-SSCP: chum salmon (Oncorhynchus keta). Canadian validation by a collaborative study and investigation of Journal of Fisheries and Aquatic Sciences 55, 1748–58. intra-species variability of the DNA patterns. Food Seeb, J.E., Kruse, G.H., Seeb, L.W. and Week, R.G. (1990) Chemistry 64, 263–8. Genetic structure of red king crab populations in Reichhardt, T. (2000) Will souped up salmon sink or Alaska facilitates enforcement of fishing regulations. swim? Nature 406, 10–12. Alaska Sea Grant College Progress Report 90-04, pp. Reisenbichler, R.R. and Rubin, S.P. (1999) Genetic 491–502. changes from artificial propagation of Pacific salmon Seeb, L.W. (1998) Gene flow and introgression within and affect the productivity and viability of supplemented among three species of rockfishes, Sebastes auricula- populations. ICES Journal of Marine Science 56, tus, S. cauriner, and S. maliger. Journal of Heredity 89, 459–66. 393–403. Richardson, B.J., Baverstock, P.R. and Adams, M. (1986) Seeb, L.W. and Crane, P.A. (1999a) Allozymes and Allozyme Electrophoresis: A Handbook for Animal mitochondrial DNA discriminate Asian and North Systematics and Population Studies. Sydney: Acade- American populations of chum salmon in mixed-stock mic Press. fisheries along the south coast of the Alaska peninsula. Romer, U. and Beisenherz, W. (1996) Environmental de- Transactions of the American Fisheries Society 128, termination of sex in Apistogramma (Cichlidae) and 88–103. two other freshwater fishes (Teleostei). Journal of Fish Seeb, L.W. and Crane, P.A. (1999b) High genetic hetero- Biology 48, 714–25. geneity in chum salmon in western Alaska, the con- Rousset, F. and Raymond, M. (1997) Statistical analyses tact zone between northern and southern lineages of population genetic data: new tools, old concepts. Transactions of the American Fisheries Society 128, Trends in Ecology and Evolution 12, 313–17. 58–87. Ruzzante, D.E. (1998) A comparison of several measures Seeb, L.W., Seeb, J.E. and Polovina, J.J. (1990) Genetic of genetic distance and population structure with variation in highly exploited spiny lobster Panulirus microsatellite data: bias and sampling variance. marginatus populations from the Hawaiian archipel- Canadian Journal of Fisheries and Aquatic Sciences ago. Fisheries Bulletin USA 88, 713–18. 55, 1–14. Shaklee, J.B. and Bentzen, P. (1998) Genetic identifica- Ruzzante, D.E., Taggart, C.T. and Cook, D. (1999) A tion of stocks of marine fish and shellfish. Bulletin of review of the evidence for genetic structure of cod Marine Science 62, 589–621. (Gadus morhua) populations in the NW Atlantic and Shaklee, J.B. and Keenan, C.P. (1986) A Practical population affinities of larval cod off Newfoundland Laboratory Guide to the Techniques and Methodolo- and the Gulf of St. Lawrence. Fisheries Research 43, gy of Electrophoresis and its Application to Fish Fillet 79–97. Identification. CSIRO Marine Laboratory Report 177, Ryman, N., Utter, F. and Laikre, L. (1995) Protection of in- Hobart, Australia. traspecific biodiversity of exploited fishes. Reviews in Shaklee, J.B., Beacham, T.D., Seeb, L. and White, B.A. Fish Biology and Fisheries 5, 417–46. (1999) Managing fisheries using genetic data: case Schlupp, I. and Ryan, M.J. (1996) Mixed-species shoals studies from four species of Pacific salmon. Fisheries and the maintenance of a sexual–asexual mating sys- Research 43, 45–78. tem in mollies. Animal Behaviour 52, 885–90. Shaw, P.W., Turan, C., Wright, J.M., O’Connell, M. and Schnabel, A., Beerli, P., Estoup, A. and Hillis, D. (1998) A Carvalho, G.R. (1999) Microsatellite DNA analysis of guide to software packages for data analysis in molecu- population structure in Atlantic herring (Clupea lar ecology. In: G.R. Carvalho (ed.) Advances in Molec- harengus), with direct comparison to allozyme and ular Ecology. Amsterdam: IOS Press, pp. 291–303. mtDNA RFLP analyses. Heredity 83, 490–9. 224 Chapter 9

Sick, K. (1961) Haemoglobin polymorphism in fishes. Waples, R.S. (1998) Separating the wheat from the chaff: Nature 192, 894–6. patterns of genetic differentiation in high gene flow Silberman, J.D. and Walsh, P.J. (1992) Species identifica- species. Journal of Heredity 98, 438–50. tion of spiny lobster phyllosome larvae via ribosomal Ward, R.D. and Elliott, N.G. (1993) Heterozygosity and DNA analysis. Molecular Marine Biology and morphological variability in the orange roughy, Biotechnology 1, 195–205. Hoplostethus atlanticus (Teleostei: Trachichthyidae). Slatkin, M. (1995) A measure of population subdivision Canadian Journal of Fisheries and Aquatic Sciences based on microsatellite allele frequencies. Genetics 50, 1641–9. 139, 157–62. Ward, R.D. and Elliott, N.G. (2001) Genetic population Smith, P.J. and Benson, P.G. (1997) Genetic diversity in structure of species in the South East Fishery of Aus- orange roughy from the east of New Zealand. Fisheries tralia. Marine and Freshwater Research 52, 563–73. Research 31, 197–213. Ward, R.D. and Grewe, P.M. (1994) Appraisal of molecu- Smith, P.J., Jamieson, A. and Bishop, C.A. (1989) Mito- lar genetic techniques in fisheries. Reviews in Fish Bi- chondrial DNA in the Atlantic cod, Gadus morhua: ology and Fisheries 4, 300–25. lack of genetic divergence between eastern and west- Ward, R.D., Skibinski, D.O.F. and Woodwark, M. (1992) ern populations. Journal of Fish Biology 34, 369–73. Protein heterozygosity, protein structure and taxo- Smith, P.J., Francis, R.I.C.C. and McVeagh, M. (1991) nomic differentiation. Evolutionary Biology 26, Loss of genetic diversity due to fishing pressure. Fish- 73–159. eries Research 10, 309–16. Ward, R.D., Woodwark, M. and Skibinski, D.O.F. (1994) Strussmann, C.A., Saito, T., Usui, M., Yamada, H. and A comparison of genetic diversity levels in marine, Takashima, F. (1997) Thermal thresholds and critical freshwater and anadromous fish. Journal of Fish Biol- period of thermolabile sex determination in two ogy 44, 213–32. atherinid fishes, Odontesthes bonariensis and Patago- Ward, R.D., Elliott, N.G. and Grewe, P.M. (1995) Al- nina hatcheri. Journal of Experimental Zoology 278, lozyme and mitochondrial DNA separation of Pacific 167–77. northern bluefin tuna, Thunnus thynnus orientalis Swearer, S.E., Caselle, J.E., Lea, D.W. and Warner, R.R. (Temminck and Schlegel), from southern bluefin tuna, (1999) Larval retention and recruitment in an island Thunnus maccoyii (Castelnau). Marine and Freshwa- population of a coral-reef fish. Nature 402, 799–802. ter Research 46, 921–30. Talaat, A.M., Reimschuessel, R. and Trucksis, M. (1997) Ward, R.D., Elliott, N.G., Innes, B.H., Smolenski, A.J. Identification of mycobacteria infecting fish to the and Grewe, P.M. (1997) Global population structure of species level using polymerase chain reaction and yellowfin tuna (Thunnus albacares) inferred from restriction enzyme analysis. Veterinary Microbiology allozyme and mitochondrial DNA variation. Fisheries 58, 229–37. Bulletin USA 95, 566–75. Tautz, D. (1989) Hypervariability of simple sequences as Welsh, J. and McClelland, M. (1990) Fingerprinting a good general source for polymorphic DNA markers. genomes using PCR with arbitrary primers. Nucleic Nucleic Acids Research 17, 6463–71. Acids Research 18, 7213–18. Tave, D. (1993) Genetics for Fish Hatchery Managers. Whitmore, D.H. (ed.) (1990) Electrophoretic and Isoelec- New York: Van Nostrand Reinhold. tric Focusing Techniques in Fisheries Management. Van Eenennaam, A.L., Van Eenennaam, J.P., Medrano, J.F. Boca Raton, FL: CRC Press. and Doroshov, S.I. (1999) Evidence of female heteroga- Williams, J.G.K., Kubelik, A.R., Livak, K.J., Rafalski, J.A. metic genetic sex determination in white sturgeon. and Tingey, S.V. (1990) DNA polymorphisms ampli- Journal of Heredity 90, 231–3. fied by arbitrary primers are useful as genetic markers. Verspoor, E. and Hammar, J. (1991) Introgressive hy- Nucleic Acids Research 18, 6531–5. bridization in fishes: the biochemical evidence Journal Wilson, C.C. and Bernatchez, L. (1998) The ghost of hy- of Fish Biology 39 (Suppl. A), 309–34. brids past: fixation of arctic charr (Salvelinus alpinus) Vos, P., Hogers, R., Bleeker, M. et al. (1995) AFLP: a new mitochondrial DNA in an introgressed population technique for DNA fingerprinting. Nucleic Acids Re- of lake trout (S. namaycush). Molecular Ecology 7, search 23, 4407–14. 127–32. Vrijenhoek, R.C. (1994) Unisexual fish: model systems Wright, S. (1969) Evolution and the Genetics of for studying ecology and evolution. Annual Review of Populations, Vol. 2. The Theory of Gene Frequencies. Ecology and Systematics 25, 71–96. Chicago: University of Chicago Press. Waples, R.S. (1987) A multispecies approach to the analy- Yearsley, G.K., Last, P.R. and Ward, R.D. (1999) sis of gene flow in marine shore fishes. Evolution 41, Australian Seafood Handbook: Domestic Species. 385–400. Hobart, Australia: CSIRO Marine Research. 10 Behavioural Ecology of Reproduction in Fish

ELISABET FORSGREN, JOHN D. REYNOLDS AND ANDERS BERGLUND

10.1 GENERAL behaviour and responses of populations to human INTRODUCTION exploitation. Our review is mainly restricted to teleost fishes and is biased towards species with Fish exhibit an enormous diversity of reproductive parental care, for the simple reason that most of behaviour. Indeed, this diversity far exceeds that the present-day knowledge on these issues comes found in birds, mammals, reptiles or amphibians. from studies of such species. Our chapter comple- For example, forms of parental care range from ments many of the others in this volume, espe- none at all in the majority of fish species, to live- cially the review of life histories by Hutchings bearing, mouth brooding or biparental guarding (Chapter 7, this volume), which covers alterna- of free-swimming young. Courtship behaviour is tive life histories stemming from reproductive also diverse, with either males or females being the conflict. dominant competitor when attracting the oppo- site sex. Social behaviour also varies greatly, from extended monogamy to promiscuity, from co- 10.2 INTRODUCTION TO operative breeding to solitary nesting, from group BREEDING SYSTEMS spawning to pair spawning. Finally, fish also ex- hibit at least five major forms of hermaphroditism, The term ‘breeding system’ (or mating system) including self-fertilization, which is unique is used when describing an animal’s mating among vertebrates. behaviour and parental care (Emlen and Oring In this chapter we attempt to make sense of this 1977; Reynolds 1996). It applies to groups of diversity in reproductive behaviour by using gen- animals, including populations or species. eral concepts from the field of behavioural ecology. Breeding system classification focuses on varia- This approach focuses on the individual, examin- tion in various kinds of behaviour. These typically ing costs and benefits resulting from interactions include: with other individuals and the environment in 1 form and duration of parental care by each sex; order to make inferences about evolution. First, we 2 form and duration of any pair bonds; give a brief theoretical background to the under- 3 number of mates (both ‘genetic’ and ‘social’); standing of breeding systems. Then we use this to 4 forms of courtship, coercion and competition explain the diversity of parental care and sexual se- (including competition among sperm from dif- lection in fishes. The latter emphasizes differences ferent males); between the sexes in courtship roles, competitive- 5 breeding resources defended and offered; ness, mate choice and numbers of mates. This 6 extent of mate choice (including choice among leads to the last section on links between breeding sperm from different males). 226 Chapter 10

10.2.1 A general framework sex (reviewed in Kvarnemo and Ahnesjö 1996). The more females a male can fertilize, the higher The fundamental difference between the sexes is his reproductive success will be. In females, on the size of the gametes. Females produce few, large the other hand, reproductive success is normally eggs, whereas males produce a vast number of tiny related to access to resources and not to the num- sperm. Sperm is a rapidly renewable resource, and ber of mating partners. Thus, the lifetime repro- males of most species could presumably fertilize ductive success of males is limited by access to hundreds of females over their lifetime if given the females, while females are not limited by access chance. Females, in contrast, must devote con- to males. This difference between the sexes leads siderable time and energy to each ovum before to an imbalance in the number of individuals of fertilization. In some species, the basic difference each sex that are ready to mate, termed the ‘opera- between the sexes is further increased through tional sex ratio’ (Emlen and Oring 1977). Typically, parental care. Female mammals, for example, usu- the operational sex ratio is male-biased because ally devote more of their time and energy to care of males are recycling into the mating pool faster offspring than males do. As a result, the potential than females. Thus, there is an excess of males and rate of reproduction is much lower in females than a shortage of females willing to mate at any given males. The potential reproductive rate is defined time. This has profound implications for sexual as the number of independent offspring that could selection. It would pay males to compete for and be produced per unit time by members of one sex if court as many females as possible, whereas fe- given unlimited access to members of the opposite males can afford to be more selective of their

Potential rate of Males Females reproduction Cheap gametes, Environmental Environmental Expensive shareable care, HIGH LOW gametes, no care or no care constraints constraints

Intense mating OSR Weak mating competition male-biased competition

Less selective More selective in mate choice in mate choice

Aggregation of Sex ratios resources and mates

Environment

Fig. 10.1 Framework for understanding breeding systems, based on a ‘typical’ fish where females do not provide parental care. The upper boxes under each sex indicate time or energy devoted to offspring (gametic development and parental care). Differences between the sexes cause a disparity in potential reproductive rates, which in turn leads to a male-biased operational sex ratio (OSR). The operational sex ratio affects mating competition and choosiness (lower boxes). Thus, sexual selection acts more strongly on males. The environment can affect potential reproductive rates and operational sex ratios. (Source: after Reynolds 1996; reproduced by permission of Elsevier Science.) Behavioural Ecology of Reproduction 227 mates. A schematic view of this is shown in Fig. much faster potential reproductive rate in warm 10.1, which exemplifies the breeding system of a water when egg development is faster, while the fe- typical fish. In fish, males are most often the sex male potential reproductive rate is not so strongly providing parental care (see Section 10.3). How- affected by temperature (Kvarnemo 1994). Thus, ever, they can often care for several clutches warmer water ‘causes’ more fighting in male simultaneously, retaining a higher potential repro- gobies (Kvarnemo 1998). ductive rate than females. Thus, despite providing Another important component of the environ- parental care, males still have much to gain from ment is the spatial and temporal distribution competing with each other for mates and courting of resources critical for breeding. If these are females. economically defendable, there are greater oppor- tunities for polygyny (Emlen and Oring 1977). For 10.2.2 Role of the environment example, if large numbers of female bluehead wrasse, Thalassoma bifasciatum, arrive at a The environment can influence mating systems spawning site on a coral reef, the dominant male is and sexual selection in several ways. It may, for unable to defend all females and additional males example, influence the costs and benefits of time participate in spawnings (Warner and Hoffman and energy devoted to different activities. Many 1980). Likewise, in the Japanese medaka, Oryzias studies have shown that predation pressure can latipes, the highest variation in male mating suc- reduce the intensity of sexual selection, either by cess was found when females arrived asynchro- relaxing mate choice (e.g. Forsgren 1992) or by in- nously (Grant et al. 1995). Similarly, if resources creasing the costs of attracting a mate. This shifts such as nest sites are spatially aggregated and the optima for ornamental traits and courtship therefore can be defended by a single male, this towards less conspicuous forms, exemplified by male can attract more mates than if the nesting studies of Trinidadian guppies, Poecilia reticulata sites were widely dispersed and hence impossible (reviewed in Endler 1995; Houde 1997). Predation to monopolize. This has been shown in experi- pressure may also play a key role in the provision of ments with the sand goby, a species which exhibits parental care. It may increase the benefits to the male parental care in defended nest sites. Polygyny offspring of receiving parental care, as well as the and the potential for sexual selection were highest costs to the parents of providing it. In the bi- when nest sites were large and females arrived parental convict cichlid, Cichlasoma nigrofascia- asynchronously (Lindström and Seppä 1996). In tum, males sometimes desert their broods and this the same species, nest abundance has implications happens most often at sites with the lowest brood for the mode of sexual selection. In areas with few predation pressure (Wisenden 1994). Environmen- nest sites, male competition is strong and the po- tal factors can also affect the operational sex ratio, tential for female choice is low, while the opposite which in turn can have strong effects on mating is true in areas with a high availability of nest sites competition and mate choice. For example, differ- (Forsgren et al. 1996a). In a field study of the slimy ential mortality rates between the sexes can bias sculpin, Cottus cognatus, Mousseau and Collins the operational sex ratio towards the sex least (1987) found that the abundance of nest substrates prone to predation. Males of the viviparous Amar- in lakes seemed to affect whether a population illo fish, Girardinichthys multiradiatus, are more was monogamous (many sites) or polygynous (few often caught by snakes than are females, biasing sites). The importance of ‘environmental con- the sex ratio towards females (Macías Garcia et al. straints’ is indicated in two places in Fig. 10.1; en- 1998). The operational sex ratio can also be affect- vironmental factors may affect each sex’s potential ed by temperature through differential effects on rate of reproduction or may directly affect the the potential reproductive rates of the two sexes operational sex ratio. This can have profound (Kvarnemo 1994; Ahnesjö 1995). Egg-guarding implications for sexual selection. male sand gobies, Pomatoschistus minutus, have a 228 Chapter 10

and Wootton 1995; Sargent 1997). Care-giving may 10.3 PARENTAL CARE increase predation risk due to conspicuousness or reduced agility. In a pipefish, Nerophis ophidion, Parental care can be defined as any behaviour that where males brood their young on their body, increases the fitness, defined in terms of develop- parental males were much more likely to be preyed ment and survival, of an animal’s offspring (see upon than non-caring males (Svensson 1988). Car- Clutton-Brock 1991). This should be in the inter- ing can also be energetically costly or reduce feed- est of the parent but, as we shall see, engaging in ing, leading to deterioration in body condition and parental care may not be as straightforward as one thereby increasing the risk of death through dis- might think. ease or starvation. In the river bullhead, Cottus gobio, body condition of parental males declined 10.3.1 Diversity of care in fish over the breeding season and this may have ac- counted for a peak in male mortality observed Fish vary substantially in their parental care. A during the second part of the breeding season complete lack of care is most common, followed (Marconato et al. 1993). Caring may also be costly by paternal care, biparental care and maternal care by extending the time to next breeding or by (Blumer 1982). Parental care occurs in about 20% reducing future fecundity. In an experiment with of all teleost families and is widely distributed phy- Galilee St Peter’s fish, Sarotherodon galilaeus, a logenetically. However, it is more common among biparental mouth brooder, parental care has been freshwater than marine families (Blumer 1982). shown to decrease growth and increase the num- Among fish with parental care, male care is most ber of days until the next reproductive cycle in common, occurring in two-thirds of the families both males and females (Balshine-Earn 1995). In (Gross and Shine 1981). The forms of care provided addition, care also decreased subsequent fecundity by fish are also diverse, from defence of eggs to in females. Care can also be costly to males in more advanced forms such as guarding and feeding terms of lost mating opportunities (see Section of the larvae after hatching. The most common 10.3.3). form of parental care is brooding the eggs on a sub- Parents may sometimes consume their proge- stratum. In this case, care may consist of actions ny, which is termed ‘filial cannibalism’ (Rohwer directed at the eggs, such as fanning them to pro- 1978) and is especially prevalent in fish (Dominey vide oxygen, and defence of the eggs against preda- and Blumer 1984). Consumption may include the tors. Other types of care are mouth brooding and whole clutch or only part of it. Filial cannibalism carrying the eggs, externally or internally. Com- in fish with parental care is now viewed as an prehensive overviews of parental care in fish have adaptive behaviour that maximizes the parent’s been provided by Blumer (1982) and Gross and lifetime reproductive success, instead of as an Sargent (1985). abnormal behaviour (FitzGerald 1992). Caring males may compensate for the costs of care by can- 10.3.2 Costs of care and nibalizing part of their progeny in order to remain filial cannibalism in sufficient physical condition to complete care of the current brood (Marconato et al. 1993) or to Parental care is often costly in terms of reducing be able to care for future broods (Rohwer 1978). the parent’s residual reproductive value, defined as Guardian males may also eat their entire clutches, its expected future reproductive success. These fit- especially when these are smaller than usual and ness costs are difficult to measure (Clutton-Brock the value of the clutch is lower than the costs of 1991) but can sometimes be approximated by mea- care (Petersen and Marchetti 1989). In a feeding ex- surements of time and energy devoted to care. In periment with the common goby, Pomatoschistus fish, there are a number of ways in which parental microps, which exhibits paternal care, whole care can be costly to the parent (reviewed in Smith clutch cannibalism was associated with unusually Behavioural Ecology of Reproduction 229 small clutches, while partial clutch cannibalism fore, the cost for a female of providing care may be was associated with a low feeding regime considerable in terms of lowered future fecundity. (Kvarnemo et al. 1998). Cannibalism by guarding In males, reproductive success is usually not males may generate conflicts of interests between equally strongly related to size. Thus, males would the sexes, and adds another factor that females not pay such a high ‘fecundity cost’ as females by should take into account when choosing a mate. providing care (Sargent and Gross 1993). If, how- ever, males lose opportunities for additional mat- 10.3.3 Who cares? ings by providing care, this could be very costly for them. However, in substrate-spawning fish with Which sex should provide care? This depends on egg guarding, males can usually care for clutches sex-specific costs and benefits of care. If parental from several females simultaneously. Thus, males care can be provided by a single parent alone, both are able to court and attract multiple mates and sexes would gain the same benefits from being the provide care at the same time. Important in this sole care-giver, provided that the male has not been context is that guarding of eggs is often ‘shareable’, cuckolded. However, the costs of providing care meaning that it is no more energetically demand- may differ between the sexes, and this difference ing to defend a nest containing thousands of eggs may explain why male care is common in fish than a nest containing only a few eggs. In other taxa (Sargent and Gross 1993). The costs of care are like mammals and birds, the investment in off- related to the options available to each sex if it did spring is substantial and less shareable among off- not engage in parental care. In fish, female fecun- spring. This may explain why uniparental male dity is often closely related to body size (Hutch- care is much more common in fish than in birds ings, Chapter 7, this volume). Thus, it is important and mammals. Furthermore, females in many for females to allocate resources to growth. In addi- species of fish prefer nests containing other fe- tion, females have to produce costly eggs. There- males’ eggs (reviewed by Reynolds and Jones 1999).

(a) Biparental (b) Biparental 2.7 40%

Male + Male + Female + 21–30 1 biparental female + biparental 1.8 biparental 0.6 0–10 1.4

Male 0 Female Male Male + Female 0.6% 60% 10.1 female 1.4 0.6

Male + Female + no care no care 1.6 0.2 No care 79.4

Fig. 10.2 (a) Hypothetical evolutionary transitions (arrows) between parental care states in teleost fishes. Numbers in boxes show the percentages of the 422 families surveyed by Blumer (1982) that have various forms of care. Circles show families that exhibit more than one form of care. (Source: after Gross and Sargent 1985; reproduced by permission.) (b) Evolutionary transitions between biparental care and female care among Cichlidae, based on 182 genera. Figures show number of transitions and arrow widths are proportional to this number. (Source: after Goodwin et al. 1998.) 230 Chapter 10

Evolutionary changes in the amount of parental lies in differences in reproductive success caused care provided by each sex have been reviewed by competition over mates. This may concern for fish and compared with other vertebrates by either number or quality of mates. Individuals Reynolds et al. (2002). Figure 10.2a shows hypothe- that carry traits that increase their reproductive sized directions of transitions for fish based on a success will successfully spread their genes to the review by Gross and Sargent (1985). It has been next generation. Members of one sex, usually the suggested that male care has most often evolved males, may compete with each other by contests from no care. Then females may, in some cases, be or other forms of rivalry. This competition may selected to join males due to selective advantages lead to the evolution of large body size, weapons of biparental care. Biparental care might be unsta- or status badges. Sexual selection may also take ble if males are then selected to desert their mates place if members of one sex, usually the females, in search of new mating opportunities, which may mate preferentially with certain individuals of the be the case in mouth brooders where multiple opposite sex. Mate choice may lead to the evolu- clutches cannot be cared for simultaneously. The tion of ornamental traits in the other sex, such as benefits and costs associated with this decision gaudy coloration, long fins and complex courtship have been explored in a game-theory framework displays. (Maynard Smith 1977), including a model devel- Darwin’s ideas about female mate choice met oped specifically for cichlid fishes (Balshine- strong resistance (reviewed by Cronin 1991), and it Earn and Earn 1997). Indeed, studies of fish have took almost 100 years until these theories were ac- provided the first experimental evidence demon- cepted. During the last decades it has been shown strating that males do desert broods in response to that mate choice is indeed common in many the availability of new mating opportunities species belonging to different taxa and that it (Keenleyside 1983; Balshine-Earn and Earn 1998). constitutes a strong selective force (reviewed by Female care, too, may be unstable through evolu- Andersson 1994). Fish have featured prominently tionary time if the benefits are outweighed by the in this research. While sexual selection has now costs of uniparental female care, expressed as de- been demonstrated conclusively, many questions creased survival and fecundity. This leads to the remain unanswered. Researchers are still asking interesting proposition that parental care may be a what specific benefits are gained through mate cyclical phenomenon, at least in species where the choice, how these are related to ornamentation costs of care are substantial and roughly equal for and courtship, how genetic variation in secondary both sexes (Gross and Sargent 1985). The hypothe- sexual traits can be maintained, and how conflicts sis that maternal care follows biparental care was between the sexes over matings have shaped be- supported by Goodwin et al. (1998), who examined haviour (Andersson 1994). evolutionary transitions in cichlid fishes (Fig. 10.2b). In cichlids biparental care seems to be the 10.4.1 Competition for mates ancestral state, and female care is the most com- mon form of uniparental care. Competition may occur over mates, spawning sites or nests. Direct fighting is one common form of competition in fish, although they have not 10.4 SEXUAL SELECTION evolved as many weapons as other vertebrates. The hooked jaws (kype) of many salmonids and the Darwin (1871) developed the theory of sexual se- large fangs of male crystal gobies, Crystallogobius lection to account for the evolution of secondary linearis, come to mind. Other weaponry like teeth, sexual characters. He was concerned about how saws and stinging rays seem not to be primarily traits such as the train of the peacock, Pavo crista- sexually selected. In fish, fighting commonly oc- tus, could evolve, as they seemed to confer no sur- curs by bites and chases, behaviours that perhaps vival advantage. The answer, Darwin suggested, do not require specialized weapons. As large body Behavioural Ecology of Reproduction 231 size may be advantageous in fights, it may be under intense sexual selection. Consequently, in many 10.4.2 Mate choice fish species males may be larger than females, in Benefits of mate choice spite of the fact that females are subjected to na- tural selection for larger body size because of Mate choice can be defined as any behaviour that fecundity advantages (Hutchings, Chapter 7, this increases the likelihood of mating with certain volume). members of the opposite sex. Theories to explain A bewildering array of threat signals have the evolution of mate preferences can be divided evolved in fish: morphological structures like the into those yielding either genetic or non-genetic orange caudal fin of the damselfish Chrysiptera benefits (reviewed in Andersson 1994). The magni- cyanea (Gronell 1989), colours like the red belly of tude of the benefits of choosiness will be propor- three-spined sticklebacks, Gasterosteus aculea- tional to the amount of variation in mate quality tus (Candolin 1999), or sounds as in the croaking of the opposite sex (Owens and Thompson 1994; gourami, Trichopsis vittata (Ladich 1998). Even Johnstone et al. 1996). electrical signals may be used in this context, as The nature of potential genetic benefits has in the electric knife fish Eigenmannia virescens been much debated. According to Fisherian (Hopkins 1974). Status signals may serve as honest models (Fisher 1930) male ornaments are merely indicators of competitive ability, even if there is no arbitrary traits that females find attractive. Orna- energetic cost of producing or displaying them be- mented males thus have a high mating success and cause individuals showing dishonestly large status selective females gain through the production of badges (‘cheaters’) will be exposed to escalated attractive sons, so long as ornament expression is fights with opponents bearing a similar-sized heritable. If female preference is also heritable, the but honest badge (e.g. Maynard Smith and Harper ornamental trait and the preference will become 1988). Information about the opponent and the en- associated in the offspring. This may lead to a self- vironment can be crucial in determining whether reinforcing runaway process causing exaggeration fighting will occur or not (e.g. Lindström 1992; of the ornamental trait. Evidence has been found Johnsson and Åkerman 1998). In fact, fights can for positive genetic correlations between female themselves provide important information on preferences and preferred male ornaments, such how to proceed (e.g. Oliveira et al. 1998). as in the three-spine stickleback (Bakker 1993). Male mating success may not translate directly However, a positive genetic correlation between into fertilization success. If females mate with preference and ornament is predicted not only more than one male, male fertilization success by Fisherian models but also by good genes models may also depend on factors like sperm number and (below). sperm quality, and select for increased sperm pro- Good genes models, also known as viability in- duction. Accordingly, across fish species there is a dicator models, suggest that male ornament ex- strong correlation between degree of sperm com- pression signals the genetic quality of the bearer. petition and relative testes size (Stockley et al. By mating with males with extravagant orna- 1997). The pipefish Syngnathus typhle, with a ments females gain indirect benefits through the paternity confidence of 100%, has minute testes, production of offspring that grow more quickly, are whereas species with intense sperm competition more viable or are more fecund (carrying ‘good like paddlefish Polyodon spathula, minnows genes’). In most of these models, the trait is Phoxinus phoxinus, silversides Menidia beryllina thought to be costly to produce or carry and only and perch Perca fluviatilis and P. flavescens have males of high quality can afford to have such a large testes and/or large sperm numbers (Stockley ‘handicap’ (Zahavi 1977). Thus, the handicap will et al. 1997). act as an honest signal. A special case of this was proposed by Hamilton and Zuk (1982), suggesting that ornamental traits reflect the bearer’s re- 232 Chapter 10 sistance to parasites. A mechanism for how this (a) 5.0 might work was suggested by Folstad and Karter (1992) in their immunocompetence handicap model: the primary androgenic hormone, testos- terone, stimulates ornament expression at the 1.5 same time as it interferes with the immune sys-

tem. Thus, only high-quality healthy males would (no. of displays) Female preference afford a high ornament expression associated with high androgen levels. Another type of genetic 0 benefit comes from mating with an individual hav- (b) 31.0 ing a very different genotype, thereby obtaining complementary genes for the offspring (Jennions 1997). Whether this occurs in fish is an open question. Wedekind (1996) speculated that this 28.5

might explain strong individual preferences (mm) among female roach.

Mating preferences may also have evolved Daughter growth through direct selection. Here, females gain direct 26.0 (non-genetic) fitness benefits that increase their survival or reproductive success. Examples of (c) 365 direct benefits are access to a safe and resource- rich territory, nuptial gifts and provision of high- quality parental care (reviewed in Andersson 1994). 280 Numerous studies of mate choice have been carried out on fish, which provide good model (mg offspring) systems that are often suitable for experimental Daughter reproduction manipulations. Many of these studies have been 215 concerned with what type of benefits females 20.0 22.5 25.0 might gain through their choice of mate. Whether Male size (mm) females are able to pick a mate with good genes is still a controversial question (Andersson 1994). Fig. 10.3 Results from an experiment designed to test One example from fish comes from an experiment for potential genetic benefits in female mate choice by on guppies, where Reynolds and Gross (1992) Trinidadian guppies. (a) Females prefer larger males (total length). (b) These same males sire faster-growing (Fig. 10.3) found that matings with preferred large daughters (total length at a standard age of 78 days). (c) males produced offspring that had a higher growth These daughters have higher reproductive output over rate and daughters with a higher reproductive suc- their first two broods (sum of dry weight at birth). Each cess. However, recent findings on the same species data point represents the mean preference or offspring suggest that another component of male attrac- performance of two females tested per male. The tiveness, their ornamentation, may be related scales are log-transformed with labels in original units. to lower offspring survival (Brooks 2000). In a (Source: from Reynolds and Gross 1992.) pipefish, both male and female choice resulted in higher offspring quality in terms of ability to es- cape predation. In addition, female choice resulted differential allocation into eggs or parental care in higher offspring growth rate (Sandvik et al. depending on the attractiveness of the partner 2000). There might, however, be other explana- (reviewed in Sheldon 2000). Several studies, for tions than good genes for these results, such as example on guppies (Houde and Torio 1992) and Behavioural Ecology of Reproduction 233 sticklebacks (Milinski and Bakker 1990, but see phus breviceps, heavily parasitized females were Fitzgerald et al. 1993), have shown that fish avoid less selective than less parasitized females, appar- parasitized mates. This preference might render ently because they had depleted energy stores and genetic benefits in terms of parasite resistance were in poor physical condition (Poulin 1994; genes passed on to the offspring (Hamilton and Barber and Poulin, Chapter 17, this volume). In Zuk 1982). However, females may also gain direct several experiments, increased predation risk has benefits by avoiding contagious parasites or by been found to result in decreased choosiness (e.g. avoiding males that are too sick to provide ade- Forsgren 1992; Berglund 1993). Although choosi- quate parental care (see also Barber and Poulin, ness apparently is very cost sensitive and may Chapter 17, this volume). disappear if search costs are high, females may be Evidence for direct benefits of mate choice willing to pay some costs to obtain preferred through high-quality parental care comes from mates. In redlip blennies, Ophioblennius atlanti- several correlational studies (e.g. Downhower and cus, some females travelled quite extensive dis- Brown 1980; Côté and Hunte 1989). Choice for tances in search of mates, enduring severe attacks parental ability has also been confirmed in experi- from damselfish whose territories they crossed in ments where confounding variables such as female the process. However, these females obtained quality and clutch size were controlled (Knapp and higher quality mates than females with a more re- Kovach 1991; Forsgren 1997a). In the study of the stricted search (Reynolds and Côté 1995). In the bicolor damselfish, Stegastes partitus, by Knapp damselfish Stegastes nigricans, females changed and Kovach (1991), females were apparently able to the number of spawning visits depending on the select good fathers by using courtship intensity as distance to the males’ territories, probably in order a cue. to minimize costs in terms of intrusions to their own territories and attacks from heterospecifics (Karino and Kuwamura 1997). Mate sampling and costs of mate choice Several models of search behaviour have been de- Ornaments and other choice cues veloped to describe how animals should search for mates (see Luttbeg 1996 and references therein). Many examples from fish show that females However, very few empirical studies on taxa other base their mate choice on male characteristics than birds have been devoted to this problem (re- (reviewed in Andersson 1994). Examples of male viewed in Gibson and Langen 1996). Female sand traits that may have evolved through female gobies appear to use a threshold criterion when se- choice include coloration (Milinski and Bakker lecting potential partners (Forsgren 1997b). Males 1990; Houde 1997), elongated fins (Bischoff et al. were inspected sequentially until one advertising 1985; Basolo 1990), body size (Côté and Hunte above a certain threshold courtship intensity was 1989; Reynolds and Gross 1992) and courtship dis- encountered and mating occurred. In the bicolour plays (Knapp and Kovach 1991). If these are honest damselfish (Knapp 1993) and the blenny Aid- signals of male quality, only high-quality males ablennius sphynx (Kraak and van den Berghe should be able to develop and carry highly elabo- 1992), females seemed to assess the success of rated versions of the traits (Zahavi 1977). Elaborate earlier laid eggs and avoided nests where eggs had traits can indeed be costly, as evidenced by the disappeared. well-studied colour ornaments of male guppies Search costs should reduce choosiness (reviewed in Endler 1995; Houde 1997; see also (Crowley et al. 1991) and several studies of fish Olson and Owens 1998 for a discussion of costs have demonstrated this. In an experiment on stick- of carotenoid colours). Also, in the Amarillo fish lebacks, females became less choosy if time costs elaborate traits are costly, as males with a mor- and energy expenditure were increased (Milinski phology attractive to females were more likely to and Bakker 1992). In the upland bullie, Gobiomor- be captured by predatory snakes (Macías Garcia 234 Chapter 10 et al. 1994). Honest signalling may work if the in Section 10.3.3, females of several different fish optimum trait expression is different for indi- prefer to spawn in nests already containing eggs viduals of different quality. This can come about from other females. Several hypotheses have been if poor-quality males pay higher costs than proposed to explain this, including mate choice high-quality males for a certain trait expression copying, a predator dilution effect and high (Johnstone 1997). To date this has not been conclu- parental investment (see Jamieson 1995). An in- sively demonstrated in any fish. There are, how- creased egg survival, resulting from the two latter ever, data supporting the idea that ornaments are mechanisms, might be a more likely explanation expressed in a condition-dependent way. For ex- than mate choice copying (Jamieson 1995; ample, parasitized males have been found to de- Forsgren et al. 1996b). velop a less bright carotenoid-based breeding It is not only males who carry extravagant orna- coloration in sticklebacks (Milinski and Bakker ments; females of many species are conspicuous in 1990) and guppies (Houde and Torio 1992). Fe- one way or another. In many fish, for example, fe- males may also gain reliable information from males are brightly coloured. The function of fe- cues that are honest by design rather than through male ornaments in species with conventional sex costs (Hasson 1997), such as body size or visible roles (see Section 10.4.3) is poorly understood in parasites. general, and most of our knowledge comes from Females may also base their choice on aspects studies on birds (Amundsen 2000). Traditionally, other than the male himself. Many fish have female ornaments have been assumed to be func- resource-based breeding systems, with males de- tionless, existing solely because of a genetic corre- fending a spawning site. Without experimental lation between male and female traits, in cases manipulations it is often difficult to assess the im- where both sexes have similar ornamentation. Al- portance of territory and male characteristics in ternatively, female ornaments may have been se- mate choice, although there are studies indicating lected for by male choice or female competition that the breeding site, in addition to male charac- (see Section 10.4.3). Recent experiments have teristics, is an important cue in mate choice (e.g. demonstrated that males are attracted to colour- Kodric-Brown 1983). This was also the conclusion ful females in both sex-role reversed pipefish from an experiment with the river bullhead (Berglund and Rosenqvist 2001) and conventional- (Bisazza and Marconato 1988). In European bitter- role gobies (Amundsen and Forsgren 2001). If ling, Rhodeus sericeus, females base their initial female ornaments have evolved through male interest in males on the male’s courtship display, choice for direct or indirect benefits, the orna- but the final spawning decision is based on the ments should provide information about female quality of spawning site that the male is defending quality. There is evidence from Arctic charr, (Candolin and Reynolds 2001). These fish spawn in Salvelinus alpinus, that the intensity of the living freshwater mussels (Unionidae) and females carotenoid-based coloration is informative of prefer mussel species that yield the highest sur- immunocompetence or health in both sexes vival to their embryos (Mills and Reynolds 2002). (Skarstein and Folstad 1996). In a removal experiment conducted in the wild, Warner (1988a) demonstrated that female blue- Interactions between male competition and head wrasse chose particular spawning sites in- female choice stead of male characteristics. Also, elaborate nest constructions may be attractive themselves (e.g. Males successful in male competition should be in Jones and Reynolds 1999; Östlund-Nilsson 2001). better condition and might also have higher viabil- An extreme form of extrabodily ornaments are ity than unsuccessful males. It is therefore gen- found in some cichlids in the African rift lakes, erally believed that females should benefit by where males construct impressive sand bowers to mating with winners of male competition (re- attract females (McKaye et al. 1990). As mentioned viewed in Berglund et al. 1996). Thus, female Behavioural Ecology of Reproduction 235 choice may be facilitated by male competition, as from a population of the peacock blenny, Salaria suggested for sticklebacks (Candolin 1999; but see pavo, where nest sites are scarce (Almada et al. Nilsson and Nilsson 2000). There is some evidence 1995). In this species, males take up nests and pro- from fish that females do prefer dominant males vide uniparental care. Accordingly, nest-holding (Järvi 1990; Bisazza and Marin 1991) and that males are hard to find and females compete in- they may gain direct benefits from such matings tensely for access to these males. Females may also (Bisazza et al. 1989). However, there is also some compete over resources necessary for reproduction evidence suggesting that females neither prefer as in the coho salmon, Oncorynchus kisutch, dominant males nor gain any direct benefits by where females compete over oviposition sites mating with these males as opposed to less (Fleming and Gross 1994). dominant ones (Forsgren 1997a). Thus, female Although most species have conventional sex choice and male competition need not work in roles, there are examples from different taxa, in- concert. In such cases, male competition may con- cluding fish, where the sex roles are reversed. Here strain female choice. Commonly, dominant males females are the predominant competitors over have a mating advantage over other males (Kodric- matings. For example, in the pipefish Syngnathus Brown 1992; but see Houde 1988). It is, however, typhle, the form and extent of male parental care often difficult to disentangle the effects of male decreased the relative potential reproductive rates competition and female choice in order to assess of males below that of females (Berglund et al. the importance of each selective mechanism inde- 1989) (Fig. 10.4). This, in turn, influenced the pendently of the other. Very little is known about operational sex ratio so that willing females were the interactions between these two sides of sexual in excess (Berglund and Rosenqvist 1993). This selection (Qvarnström and Forsgren 1998). caused males to be choosier and females more competitive (Berglund 1991). Consequently, 10.4.3 Sex roles females should be subject to stronger sexual selec- tion than males, and females more than males Sex differences in mating competition largely de- should evolve behaviours and structures aiding pend on the operational sex ratio (Kvarnemo and them in mating competition (e.g. Berglund et al. Ahnesjö 1996; see Section 10.2.1). Species where 1997). Several other fish species also exhibit sex- males are most competitive over matings are role reversal, for example the pipefish Nerophis said to have conventional sex roles (Vincent et al. ophidion (Rosenqvist 1990), the black-chinned 1992). This does not necessarily imply that male tilapia, Sarotherodon melanotheron (Balshine- mate choice and female competition over matings Earn and McAndrew 1995) and the goby Eucyclo- are absent in these species. These behaviours are, gobius newberryi (Swenson 1997). Note that however, usually less obvious and have often paternal care per se does not cause sex-role rever- been overlooked in studies of sexual selection sal: the majority of caring fish species have exclu- (Cunningham and Birkhead 1998). When male po- sive paternal care but sex roles are typically not tential reproductive rate is not so high and female reversed (Fig. 10.1). This is probably because the variation in quality is substantial, males are ex- usual forms of care, such as guarding and fanning, pected to be choosy. In fish, several studies have allow the male to accept several clutches and does demonstrated male choice for large and more not depress his potential reproductive rate below fecund females (reviewed in Andersson 1994). that of the females. Even the extreme form of pa- Female competition over males is expected if there ternal care found in pipefishes and seahorses does is a shortage of males or if variation in male quality not necessarily cause sex-role reversal: seahorses is large (reviewed in Berglund et al. 1993). In the lat- typically have conventional sex roles (Vincent et ter case females are expected to compete over the al. 1992; Vincent 1994). most preferred males, so showing competitive mate choice. An example of male shortage comes 236 Chapter 10

8

Fig. 10.4 Potential reproductive 6 rate in females of the pipefish Syngnathus typhle. Females could fill on average two males 4 during the time of an average male pregnancy (45 days) when No. of females provided with an excess of mates. 2 Thus, females have a potential reproductive rate twice as high as that of males. (Source: after 0 Berglund et al. 1989; reproduced 01234by permission of the University of No. of filled males within 45 days Chicago Press.)

10.5 MATING PATTERNS petition except for sperm competition. This may explain the lack of sexual dimorphism in this The diversity of mating patterns, ranging from species. promiscuity to monogamy, is to a large extent ex- Leks are areas where males aggregate and dis- plained by sex differences in mating competition play to females, which come there to mate only and resource distribution. Often, females distrib- (Höglund and Alatalo 1995). Females receive noth- ute themselves according to resources, such as ing but sperm from the males, and do not leave eggs food, shelter or breeding sites, while males distrib- or young on the males’ territories. A spectacular ute themselves according to female distribution. example is the bower-building cichlid, Cyrtocara Males may compete for the resources females need eusinostomus, in Lake Malawi, where up to 50000 or try to monopolize females directly (Emlen and males may display simultaneously along a 4-km Oring 1977; Reynolds 1996). As resources vary in long arena. After mating, the female, who does all space and time, mating patterns may too, depend- the brooding, leaves the lek with the eggs in her ing on economic defendability (see Section 10.2.2). mouth (McKaye et al. 1990). Cod, Gadus morhua, This may, for example, explain the occurrence of may provide another example, with aggregated ‘resource-defence polygyny’. While this scheme males courting females vocally and by fin displays explains broad mating patterns, additional infor- (Nordeide and Folstad 2000). mation about parental care and sex differences in Resource-defence polygyny is common in fish costs and benefits of mate choice and competition because, in many species, males defend a spawning are needed to understand the details of interac- site or build a nest where females lay their eggs. tions between the sexes (Reynolds 1996). This is widespread among, for example, gobies, Promiscuous group-spawning is common in sticklebacks, labrids and damselfish. Female- many pelagic fish, probably because resources defence polygyny is much rarer but has been found are widely distributed in time and space and not in two simultaneously hermaphroditic sea basses, economically defendable. Many commercial fish the barred serrano, Serranus fasciatus, and the species may fall into this category, such as herring, lantern bass, S. baldwini. Males defend groups of Clupea harengus, which gather in large numbers female-acting hermaphrodites with small and col- to spawn. There are probably few options for exer- lectively defendable home ranges (Petersen 1987). cising mate choice or engaging in intrasexual com- In S. fasciatus, groups of females form with the Behavioural Ecology of Reproduction 237 dominant individual acting as a pure male. anew. A female cannot search for a new mate while Subordinates spawn as females but retain their retaining hydrated eggs, as these must be released functional testes without using them much. How- within a short period. However, one can wonder ever, these ‘females’ may streak: a subordinate whether preparatory periods select for monogamy can hide close to a pair about to spawn and dash in or whether monogamy permits the female to to release sperm at the right time (Petersen 1990). have preparatory periods. Whatever the reason, For a review of alternative mating tactics see seahorses and monogamous pipefish seem to be Hutchings (Chapter 7, this volume). examples of animals where ‘pure’ genetic mono- A few fish species are monogamous. In the gamy can be found, as evidenced by microsatellite Midas cichlid, Cichlasoma citrinellum, parents analysis in the Western Australian seahorse H. stay together to defend their territory both against angustus (Jones et al. 1998). other Midas cichlids and to protect their larvae Communal breeding has evolved in several from predators, including conspecifics. A single fish species with parental care and is especially parent is unlikely to be able to accomplish this suc- common in the Cyprinidae and the Cichlidae cessfully (e.g. Keenleyside 1991). Monogamy is a (Taborsky 1994; Wisenden 1999). Two common fixed trait in this species: even if the sex ratio and forms are ‘non-reproducing’ helpers that assist a density are experimentally manipulated to pro- reproducing couple in raising their offspring, and mote polygyny, monogamy reigns (Rogers 1987). parental fish that adopt and raise other individuals’ Biparental care is, however, not the only factor offspring (alloparents). Several advantages have that promotes monogamy. Males may be kept been proposed to explain ‘helping’. If helpers are from access to several females by dominant males related to the individuals they assist, kin selection and may have to resort to monogamy or not mate may explain this phenomenon. Thus helpers at all, as in the cichlid Lamprologus brichardi propagate their own genes by helping relatives, as (Limberger 1983). In marine fishes, monogamy is in the cichlid Lamprologus brichardi (Taborsky commonly associated with difficulties in finding and Limberger 1981). Helpers may also benefit by and keeping more than one mate. Monogamy may being included in a group and gain access to shel- be influenced by territoriality, low population ters or feeding sites if such are scarce (Taborsky density, and restricted or slow movement. In two 1984). Further, helpers may reproduce by stealing species of coral-dwelling hawkfishes, Neocirrhites fertilizations from the territorial male, as in the armatus and Oxycirrhites typus, males were cichlid Neolamprologus pulcher (Dierkes et al. monogamous if suitable corals were small or rare, 1999). Alloparents may benefit if larger schools of but maintained groups of females among large offspring reduce the per-capita predation by dilut- or abundant corals (Donaldson 1989). In the ing the risk or by confusing the predator (reviewed simultaneous hermaphrodite Serranus fasciatus, in Pitcher and Parrish 1993). In addition, predation pairs spawn monogamously under low population risk may be diverted from the parent’s young to the densities, taking turns to act as males and females alien young, as in convict cichlids (Wisenden and (Petersen 1990). Keenleyside 1993). However, alloparenting need Reproductive synchrony between males and fe- not always be adaptive but can simply be mis- males may select for monogamy. In the pipefish directed care, as when mouth-brooding cichlids Corythoichthys intestinalis (Gronell 1984) and in are parasitized by the catfish Synodontis multi- two seahorses, Hippocampus fuscus and H. whitei punctatus (Sato 1986). (Vincent et al. 1992; Vincent and Sadler 1995), Some species of fish do not have sex at all. females need to hydrate their eggs for some time Examples include some minnows of the genus before depositing them in the male brood pouch. Phoxinus and the topminnow genus Poeciliopsis. Under monogamy, the female hydrates eggs while The latter contains both sexual and asexual the male broods the previous clutch. Once the species. The asexual all-female species were male has given birth, the pair is ready to mate formed by hybridization between two sexual top- 238 Chapter 10 minnow species and are triploid. Sperm from one driven either by sexual selection or by natural of the ancestral species is still required to initiate selection for fecundity advantages. reproduction but is not incorporated into the off- Protogyny, in which individuals start as spring (Vrijenhoek 1979). Males of the ancestral females and then switch to males, is the common- species still benefit from spawning with the asexu- est form of hermaphroditism, exemplified by al females because females of their own species groupers (Epinephilinae), parrotfishes (Scaridae) copy the mate choice of the asexual females and many wrasses (Labridae). This sexually select- (Schlupp et al. 1994). These fish are often infected ed life-history trait is explained by the ‘size-advan- by trematode larvae (Uvulifer sp.) and may develop tage model’ (Fig. 10.5; Ghiselin 1969; reviewed in the black spot disease. Theory attempting to ex- Warner 1988b). In species with strong competition plain why sex is advantageous suggests that asexu- als should be more parasitized than sexuals, and that variation in parasite level should be higher in (a) sexuals. This has been confirmed for an assem- Polygyny potential high blage of one sexual and two asexual topminnow species (Lively et al. 1990). Further, when the sexual species was inbred following a popula- tion bottleneck, parasite levels rose, suggesting that a lowered genetic variability increases susceptibility.

10.6 REPRODUCTIVE Change BEHAVIOUR AND LIFE HISTORIES

Fecundity (b) Reproductive behaviour is intimately linked with Polygyny potential low life histories because the time and energy devoted to mating behaviour and parental care usually incur costs in terms of decreased future survival and fecundity. It is a matter of semantics whether one considers parental care to be a component of life histories. Below we address briefly some areas of reproductive behaviour that have close links to life histories. Life histories are reviewed in detail Change in Hutchings, (Chapter 7, this volume). Age or size

10.6.1 Sex change Fig. 10.5 The size-advantage model of sex change. (a) and hermaphrodites When only large males can defend resources or females, it pays to start life as a female and change sex when a Many coral-reef fish and some temperate species competitive size/age is reached (protogyny). (b) When are sequential or simultaneous hermaphrodites. there is a low potential for polygyny, female fecundity Sequential hermaphrodites may change sex from usually increases more rapidly with size than male fecundity. It pays to start out as male and change to male to female, female to male, or both ways. female when reaching some size/age threshold Simultaneous hermaphrodites produce eggs and (protandry). These scenarios assume that costs of sperm at the same time and may either cross- or changing are not prohibitive. (Source: after Berglund self-fertilize. This feature of life histories could be 1997; © Oxford University Press 1997.) Behavioural Ecology of Reproduction 239 among males for females, small males may be at a behaviour called egg trading. Eggs are more strong disadvantage. Then it may pay individuals expensive to produce than sperm, so a pair benefits to start out as females and remain so until they from sharing the heavy task of egg production, reach a sufficient size to compete successfully as taking turns spawning as a male and as a female. males (Fig. 10.5a). At this point the gains from However, if a ‘female’ allows its partner to play the reproducing as a male outweigh the lower-risk, ‘cheap’ male role, how can ‘she’ know that ‘she’ lower-benefit option of reproducing as a female, can be a ‘he’ next time? In other words, how is including the time and energy costs of switch- reciprocation achieved in this obvious conflict ing over their gametic function. So they switch to between the sexes? This is a classic problem in males. Experiments have shown that when domi- studies of human social conflict, known as the nant males are removed, the largest female in an Prisoner’s dilemma (Axelrod and Hamilton 1981). assemblage may initiate sex change within hours It falls in the domain of game theory, because an and become fully functional within days (Ross individual’s best response depends on what others 1990). in the population do (see also discussion of game Protandry, in which sex changes from male to theory by Hannesson, Chapter 12, Volume 2). female, usually occurs in species where sexual Hamlets, Hypoplectrus nigricans, have solved this selection through male–male competition is less dilemma in the same way that a computer tourna- important, for instance when breeding resources ment eventually solved it, using the ‘tit-for-tat’ are less aggregated. Individuals are then selected to strategy, i.e. always start by cooperating and then switch from male to female when the benefits do what your partner did in the previous round of higher fecundity associated with larger body (Fischer 1980; Leonard 1993). Each member of the size outweigh the benefits of remaining a male pair takes turns acting as a female and as a male. By (Fig. 10.5b). This has been suggested to explain allowing a male-acting fish to fertilize only a protandry in the anemonefish Amphiprion small parcel of eggs (egg trading) and then with- akallopisos (Fricke 1979). If the female of a pair holding eggs unless the other individual recipro- is removed, the top-ranking male changes into a cates by acting female, any loss due to cheating is female and the next highest ranking male (or an reduced (Fischer 1980). immigrant if there are no more males) fertilizes There is one species of self-fertilizing her- the eggs (Fricke and Fricke 1977). maphrodite, the killifish Rivulus marmoratus Some goby species take the elements of the size- (Cyprinodontidae). Indeed, this is the world’s only advantage model one step further and switch in known self-fertilizing vertebrate, although pure both directions. In some species this is advanta- males and outcrossing may occur (Lubinski et al. geous if moving between territories is risky: after 1995). mate loss, a flexible attitude to sex increases part- ner availability and reduces mate search costs, as individuals can mate with whichever sex is avail- able (Nakashima et al. 1996; Munday et al. 1998). 10.7 REPRODUCTIVE Thus, there is no sharp borderline between sequen- BEHAVIOUR AND tial and simultaneous hermaphroditism: when EXPLOITATION changing occurs rapidly we call it simultaneous hermaphroditism (e.g. St Mary 1994). There are numerous links between the reproduc- If a male, above some certain body size, gains tive behaviour of fish and exploitation (Vincent nothing by allocating further resources to sperm and Sadovy 1998; Reynolds and Jennings 2000). For production, he may instead allocate new resources example, fishers in many parts of the world take to egg production: we now have a simultaneous advantage of predictable spawning aggregations of hermaphrodite. Male size advantages may be fish, and managers often incorporate this informa- circumvented by low mate encounter rates and a tion into their plans. Reproductive behaviours 240 Chapter 10 may affect the vulnerability of fish to capture, as ringens, the Canadian cod and Atlantoscandian well as the resilience of populations in responding herring, which have all experienced crashes to exploitation. Other forms of exploitation, such largely due to overfishing. Large species of as fish farming and sea-ranching, may also affect groupers in the Indo-Pacific may aggregate in the fish populations. Here, altered or relaxed selection thousands at traditional sites. Unrestricted fishing pressures may lead to altered behaviour, morphol- access to these sites can result in rapid depletions ogy and life history, which in turn may affect sur- (Sadovy 1993, 1994). Orange roughy, Hoplostethus vival and reproduction (e.g. Petersson and Järvi atlanticus, gather in large spawning aggregations 1997). For example, hatchery populations of fe- at underwater peaks, i.e. seamounts. Here they are male coho salmon have lost a number of secondary vulnerable to trawling, which has proved to be a sexual characters (Fig. 10.6). Obviously, loss of considerable threat since large-scale fisheries naturally and sexually selected traits due to do- began in New Zealand during 1978. Aggregated mestication can have implications for the conser- spawning behaviour, combined with a slow life vation of wild populations, when cultured fish history (average age at maturity about 32 years; escape or are released (Fleming 1994). Fenton et al. 1991), has led to rapid declines (see also Ward, Chapter 9, this volume). The elaborate 10.7.1 Timing and location of sand bowers built by male haplochromine cichlids spawning in Lake Malawi (see Sections 10.4.2 and 10.5) are destroyed by trawlers, thereby disrupting spawn- Fish species that form spawning aggregations in ing even for the surviving males (see illustrations locations accessible to fisheries are particularly in Ribbink 1987). susceptible to overexploitation (see Reynolds et In some parts of the world, fishers take advan- al., Chapter 15, Volume 2). This includes many tage of limited spawning sites by attracting fish to anadromous species such as salmonids and stur- artificial sites where they can be caught. An exam- geons that are targeted in estuaries and rivers. It ple is the use of bundles of sugar cane, palm fronds also applies to the Peruvian anchoveta, Engraulis or banana leaves by fishers in the Caribbean target-

(a) (b) 4.1 51

3.9 49

3.7 47 Colour score 3.5 45 Hooked snout length (mm) 3.3 43 01234 01234 Breeding competition

Fig. 10.6 Relationships between breeding competition and (a) intensity of red breeding coloration and (b) hooked snout length for 11 wild populations and five hatchery populations (mean ± SE) of female coho salmon. Breeding competition for the wild populations was measured as the average female population size (measured over 3–38 years) divided by the spawning capacity of the stream. For hatchery populations, breeding competition was assumed to be zero since spawning occurs artificially and independently of natural breeding competition. Correlations are significant even if only the wild populations are included. (Source: after Fleming and Gross 1989; reproduced by permission.) Behavioural Ecology of Reproduction 241 ing flying fish. The fish are caught during their (Warner et al. 1995). It remains to be seen whether frantic spawning activities, often while attempt- these results apply to species of conservation con- ing to spawn on the mesh of the nets themselves. cern such as groupers. Artificial reefs are also used to attract fish, and may Sex-specific fisheries may also result from under certain circumstances increase productivity spatial and temporal segregation of the sexes (Pickering and Whitmarsh 1997). (reviewed in Vincent and Sadovy 1998). As with considerations of sex change, this is apt to be most 10.7.2 Sex change and important where the sex that is targeted has the sex-specific fisheries lower potential reproductive rate, and hence may limit population productivity. For example, in the An understanding of behavioural ecology is essen- coral trout, Plectropomus areolatus, females tial in order to predict the effects of sex change and are caught selectively before spawning, leading to sex-specific fisheries on the resilience of popula- male sex-biases in spawning aggregations and tions to fishing pressures (Vincent and Sadovy male harassment of females that may hinder 1998; Petersen and Warner 2002). Most sex- spawning (Johannes et al. 1994). In some migratory changing reef fishes change from female to male species such as plaice, Pleuronectes platessa, when they reach a large size (Section 10.6.1). If sexual selection is probably responsible for males this switch is under social control, i.e. triggered by arriving at spawning grounds before females the loss of large males, then one might expect that (Arnold and Metcalfe 1996), leading to sex-specific the loss of males would not impair productivity. fisheries during migration. However, there are at least two reasons why we Anglers may have adverse effects on male fish should not take much comfort from this idea. that provide parental care. For example, North First, intensive fishing may not allow sufficient American male smallmouth bass, Micropterus time for females to change sex (Vincent and Sadovy dolomieu, will bite anything that comes near their 1998). This may explain why males of one species, nests, including hooks. Hook-and-release rules the gag grouper, Mycteroperca microlepis, have during the spring spawning season may allow become proportionately much rarer in the Gulf males to survive (see Cowx, Chapter 17, Volume of Mexico and the southeastern Atlantic (e.g. 2). However, an experimental study showed that McGovern et al. 1998). Second, although the social when anglers ‘played’ fish to exhaustion (2min), control of sex change is understood in some small it took the males four times longer to return to species of reef fishes that live in permanent social their nests than when they were played briefly groups, little is known about mechanisms of sex (<20s) (Kieffer et al. 1995). Even a brief absence change in large groupers that come together only of the male can lead to breeding failure due to pre- for spawning. No conclusions can be drawn about dation by other species as well as neighbouring implications of male-biased fishing for stock pro- males. ductivity until we understand better how female- biased the sex ratio needs to become before sperm becomes limiting (see Petersen and Levitan 2001; 10.8 CONCLUSIONS Petersen and Warner 2002). The evidence from small reef fishes such as the bluehead wrasse is As we have seen, fish are excellent model organ- that although group spawnings may have more isms for studying important questions in be- than ten times as much sperm released as pair havioural ecology, displaying a wide variety of spawnings (Shapiro et al. 1994), this difference complex reproductive behaviours. Today, most of does not greatly affect fertilization success our knowledge of these questions comes from (Marconato et al. 1997). However, frequently studies of shallow-water teleost fish with parental spawning male bluehead wrasse release less sperm care. Indeed, researchers have relied very heavily per mating, resulting in lower fertilization success on a handful of model species, such as sticklebacks 242 Chapter 10 and guppies. We know surprisingly little about the Axelrod, R. and Hamilton, W.D. (1981) The evolution of behavioural ecology of reproduction in fish with- cooperation. Science 211, 1390–6. out parental care. Absence of parental care is typi- Bakker, T.C.M. (1993) Positive genetic correlation between female preference and preferred male orna- cal of many pelagic fish, including some of the ment in sticklebacks. Nature 363, 255–7. most abundant and economically important fish Balshine-Earn, S. (1995) The costs of parental care in species on earth, such as cod and anchovetas. Galilee St Peter’s fish, Sarotherodon galilaeus. Animal Whether those species fit into the general picture, Behaviour 50, 1–7. for example, with widespread sexual selection, is Balshine-Earn, S. and Earn, D.J.D. (1997) An evolutionary basically unknown. An improved knowledge of model of parental care in St Peter’s fish. Journal of the behavioural and reproductive dynamics of Theoretical Biology 184, 423–31. Balshine-Earn, S. and Earn, D.J.D. (1998) On the evolu- such species, focusing on the adaptive decisions tionary pathway of parental care in mouth-brooding made by individual fish, is clearly needed. We cichlid fish. Proceedings of the Royal Society of believe that increased communication and London Series B 265, 2217–22. collaboration between behavioural ecologists, Balshine-Earn, S. and McAndrew, B.J. (1995) Sex-role fisheries biologists and nature conservationists reversal in the black-chinned tilapia, Sarotherodon would be very fruitful and could greatly increase melanotheron (Ruppel) (Cichlidae). Behaviour 132, 861–74. our knowledge and improve resource Basolo, A.L. (1990) Female preference predates the evolu- management. tion of the sword in swordtail fish. Science 250, 808–10. Berglund, A. (1991) Egg competition in a sex-role reversed ACKNOWLEDGEMENTS pipefish: subdominant females trade reproduction for growth. Evolution 45, 770–4. Berglund, A. (1993) Risky sex: male pipefish mate at ran- We thank Trond Amundsen for valuable com- dom in the presence of a predator. Animal Behaviour ments on the chapter and Hazel Oxenford for help- 46, 169–75. ful information. Berglund, A. (1997) Mating systems and sex allocation. In: J.G. Godin (ed.) Behavioural Ecology of Teleost Fishes. Oxford: Oxford University Press, pp. 237–65. Berglund, A. and Rosenqvist, G. (1993) Selective males REFERENCES and ardent females in pipefishes. Behavioral Ecology and Sociobiology 32, 331–6. Ahnesjö, I. (1995) Temperature affects male and female Berglund, A. and Rosenqvist, G. (2001) Male pipefish potential reproductive rates differently in the sex-role prefer ornamented females. Animal Behaviour 61, reversed pipefish Syngnathus typhle. Behavioral 345–50. Ecology 6, 229–33. Berglund, A., Rosenqvist, G. and Svensson, I. (1989) Almada, V.C., Gonçalves, E.J., Oliveira, R.F. and Santos, Reproductive success of females limited by males in A.J. (1995) Courting females: ecological constraints two pipefish species. American Naturalist 133, affect sex roles in a natural population of the blen- 506–16. niid fish Salaria pavo. Animal Behaviour 49, 1125–7. Berglund, A., Magnhagen, C., Bisazza, A., König, B. and Amundsen, T. (2000) Female ornaments: genetically Huntingford, F. (1993) Female–female competition correlated or sexually selected? In Y. Espmark, T. over reproduction. Behavioral Ecology 4, 184–7. Amundsen and G. Rosenqvist (eds) Animal signals. Berglund, A., Bisazza, A. and Pilastro, A. (1996) Arma- Trondheim: Tapir Academic Press, pp. 133–54. ments and ornaments: an evolutionary explanation of Amundsen, T. and Forsgren, E. (2001) Male mate choice traits of dual utility. Biological Journal of the Linnean selects for female coloration in a fish. Proceedings of Society 58, 385–99. the National Academy of Sciences USA 98, 13155–60. Berglund, A., Rosenqvist, G. and Bernet, P. (1997) Orna- Andersson, M. (1994) Sexual Selection. Princeton, NJ: mentation predicts reproductive success in female Princeton University Press. pipefish. Behavioral Ecology and Sociobiology 40, Arnold, G.P. and Metcalfe, J.D. (1996) Seasonal migra- 145–50. tions of plaice (Pleuronectes platessa) through the Bisazza, A. and Marconato, A. (1988) Female mate Dover Strait. Marine Biology 127, 151–60. choice, male–male competition and parental care in Behavioural Ecology of Reproduction 243

the river bullhead, Cottus gobio L. (Pisces, Cottidae). tion and the evolution of mating systems. Science 197, Animal Behaviour 36, 1352–60. 215–23. Bisazza, A. and Marin, G. (1991) Male size and female Endler, J.A. (1995) Multiple trait coevolution and mate choice in the eastern mosquitofish (Gambusia environmental gradients in guppies. Trends in Ecology holbrooki: Poeciliidae). Copeia 1991, 730–5. and Evolution 10, 22–9. Bisazza, A., Marconato, A. and Marin, G. (1989) Male Fenton, G.E., Short, S.A. and Ritz, D.A. (1991) Age deter- competition and female choice in Padogobius mination of orange roughy, Hoplostethus atlanticus martensi (Pisces, Gobiidae). Animal Behaviour 38, (Pisces: Trachichthyidae) using 210Pb: 226Ra disequilib- 406–13. ria. Marine Biology 109, 197–202. Bischoff, R.J., Gould, J.L. and Rubenstein, D.I. (1985) Tail Fischer, E.A. (1980) The relationship between mating size and female choice in the guppy (Poecilia reticulta). system and simultaneous hermaphroditism in the Behavioral Ecology and Sociobiology 17, 253–5. coral reef fish, Hypoplectrus nigricans (Serranidae). Blumer, L.S. (1982) A bibliography and categorization Animal Behaviour 28, 620–33. of bony fishes exhibiting parental care. Zoological Fisher, R.A. (1930) The Genetical Theory of Natural Journal of the Linnean Society 76, 1–22. Selection. Oxford: Clarendon Press. Brooks, R. (2000) Negative genetic correlation between FitzGerald, G.J. (1992) Filial cannibalism in fishes: why male sexual attractiveness and survival. Nature 406, do parents eat their offspring? Trends in Ecology and 67–70. Evolution 7, 7–10. Candolin, U. (1999) Male–male competition facilitates FitzGerald, G.J., Fournier, M. and Morrisette, J. (1993) female choice in sticklebacks. Proceedings of the Sexual selection in an anadromous population of Royal Society of London Series B 266, 785–9. threespine sticklebacks: no role for parasites. Evolu- Candolin, U. and Reynolds, J.D. (2001) Sexual signaling tionary Ecology 8, 348–56. in the European bitterling: females learn the truth by Fleming, I.A. (1994) Captive breeding and conservation inspecting the resource. Behavioral Ecology 12, of wild salmon populations. Conservation Biology 8, 407–11. 886–8. Clutton-Brock, T.H. (1991) The Evolution of Parental Fleming, I.A. and Gross, M.R. (1989) Evolution of Care. Princeton, NJ: Princeton University Press. adult female life history and morphology in a pacific Côté, I.M. and Hunte, W. (1989) Male and female mate salmon (coho: Onchorynchus kisutch). Evolution 43, choice in the redlip blenny: why big is better. Animal 141–57. Behaviour 38, 78–88. Fleming, I.A. and Gross, M.R. (1994) Breeding competi- Cronin, H. (1991) The Ant and the Peacock. Cambridge: tion in a pacific salmon (coho: Onchorynchus kisutch): Cambridge University Press. measures of natural and sexual selection. Evolution Crowley, P., Travers, S., Linton, M., Cohn, S., Sih, A. and 48, 637–57. Sargent, R.C. (1991) Mate density, predation risk, and Folstad, I. and Karter, A.J. (1992) Parasites, bright males the seasonal sequence of mate choices: a dynamic and the immunocompetence handicap. American game. American Naturalist 137, 567–96. Naturalist 139, 603–22. Cunningham, E.J.A. and Birkhead, T.R. (1998) Sex roles Forsgren, E. (1992) Predation risk affects mate choice in a and sexual selection. Animal Behaviour 56, 1311–21. gobiid fish. American Naturalist 140, 1041–9. Darwin, C. (1871) The Descent of Man, and Selection in Forsgren, E. (1997a) Female sand gobies prefer good Relation to Sex. London: Murray. fathers over dominant males. Proceedings of the Royal Dierkes, P., Taborsky, M. and Kohler, U. (1999) Reproduc- Society of London Series B 264, 1283–6. tive parasitism of broodcare helpers in a cooperatively Forsgren, E. (1997b) Mate sampling in a population of breeding fish. Behavioral Ecology 10, 510–15. sand gobies. Animal Behaviour 53, 267–76. Dominey, W.J. and Blumer, L.S. (1984) Cannibalism of Forsgren E., Kvarnemo, C. and Lindström, K. (1996a) early life history stages in fishes. In: G. Hausfater and Mode of sexual selection determined by resource S.B. Hrdy (eds) Infanticide: Comparative and Evolu- abundance in two sand goby populations. Evolution tionary Perspectives. New York: Aldine, pp. 43–64. 50, 646–54. Donaldson, T.J. (1989) Facultative monogamy in obligate Forsgren E., Karlsson, A. and Kvarnemo, C. (1996b) coral-dwelling hawkfishes (Cirrhitidae). Environmen- Female sand gobies gain direct benefits by choosing tal Biology of Fishes 26, 295–302. males with eggs in their nest. Behavioral Ecology and Downhower, J.F. and Brown, L. (1980) Mate preferences Sociobiology 39, 91–6. of female mottled sculpins Cottus bairdi. Animal Fricke, H.W. (1979) Mating system, resource defence and Behaviour 28, 728–34. sex change in the anemonefish Amphiprion akallop- Emlen, S.T. and Oring, L.W. (1977) Ecology, sexual selec- isos. Zeitschrift für Tierpsychologie 50, 313–26. 244 Chapter 10

Fricke, H.W. and Fricke, S. (1977) Monogamy and sex Johannes, R.E., Squire, L. and Graham, T. (1994) Develop- change by aggressive dominance in coral reef fish. ing a protocol for monitoring spawning aggregations Nature 266, 830–2. of Palauan serranids to facilitate the formulation and Ghiselin, M.T. (1969) The evolution of hermaphroditism evaluation of strategies for their management. South among animals. Quarterly Review of Biology 44, Pacific Forum Fisheries Agency Report 94/28. 189–208. Honiara, Solomon Islands. Gibson, R.M. and Langen, T.A. (1996) How do animals Johnsson, J.I. and Åkerman, A. (1998) Watch and learn: choose their mates? Trends in Ecology and Evolution preview of the fighting ability of opponents alters 11, 468–70. contest behaviour in rainbow trout. Animal Behaviour Goodwin, N.B., Balsine-Earn, S. and Reynolds, J.D. (1998) 56, 771–6. Evolutionary transitions in parental care in cichlid Johnstone, R.A. (1997) The evolution of animal signals. fish. Proceedings of the Royal Society of London Series In: J.R. Krebs and N.B. Davies (eds) Behavioural Eco- B 265, 2265–72. logy, 4th edn. Oxford: Blackwell Science, pp. 155–78. Grant, J.W.A., Bryant, M.J. and Soos, C.E. (1995) Opera- Johnstone, R.A., Reynolds, J.D. and Deutsch, J.C. (1996) tional sex ratio, mediated by synchrony of female Mutual mate choice and sex differences in choosiness. arrival, alters the variance of male mating success in Evolution 50, 1382–91. Japanese medaka. Animal Behaviour 49, 367–75. Jones, A.G., Kvarnemo, C., Moore, G.I., Simmons, L.W. Gronell, A.M. (1984) Courtship, spawning and social or- and Avise, J.C. (1998) Microsatellite evidence for ganization of the pipefish, Corythoichthys intestinalis monogamy and sex-biased recombination in the (Pisces: Syngnathidae) with notes on two congeneric Western Australian seahorse Hippocampus angustus. species. Zeitschrift für Tierpsychologie 65, 1–24. Molecular Ecology 7, 1497–505. Gronell, A.M. (1989) Visiting behaviour by females of the Jones, J.C. and Reynolds, J.D. (1999) The influence of oxy- sexually dichromatic angel fish, Chrysiptera cyanea gen stress on female choice for male nest structure in (Teleostei, Pomacentridae) – a probable method of the common goby. Animal Behaviour 57, 189–96. assessing male quality. Ethology 81, 89–122. Järvi, T. (1990) The effects of male dominance, secondary Gross, M.R. and Sargent, R.C. (1985) The evolution of sexual characteristics and female mate choice on the male and female parental care in fishes. American mating success of male Atlantic salmon Salmo salar. Zoologist 25, 807–22. Ethology 84, 123–32. Gross, M.R. and Shine, R. (1981) Parental care and mode Karino, K. and Kuwamura, T. (1997) Plasticity in spawn- of fertilization in ectothermic vertebrates. Evolution ing visits of female damselfish, Stegastes nigricans: 35, 775–93. effect of distance to mates. Behavioral Ecology and Hamilton, W.D. and Zuk, M. (1982) Heritable true fit- Sociobiology 41, 55–9. ness and bright birds: a role for parasites? Science 218, Keenleyside, M.H.A. (1983) Mate desertion in relation to 384–7. adult sex ratio in the biparental cichlid fish Herotilapia Hasson, O. (1997) Towards a general theory of biological multispinosa. Animal Behaviour 31, 683–8. signaling. Journal of Theoretical Biology 185, 139–56. Keenleyside, M.H.A. (1991) Parental care. In: M.H.A. Hopkins, C.D. (1974) Electric communication: function Keenleyside (ed.) Cichlid Fishes: Behaviour, Ecology in the social behavior of Eingemannia virescens. Be- and Evolution. London: Chapman & Hall, pp. 191– haviour 50, 270–305. 208. Houde, A.E. (1988) The effect of female choice and Kieffer, J.D., Kubacki, M.R., Phelan, F.J.S., Philipp, D.P. male–male competition on the mating success of male and Tufts, B.L. (1995) Effects of catch-and-release an- guppies. Animal Behaviour 36, 888–96. gling on nesting male smallmouth bass. Transactions Houde, A.E. (1997) Sex, Color, and Mate Choice in Gup- of the American Fisheries Society 124, 70–6. pies. Princeton, NJ: Princeton University Press. Knapp, R.A. (1993) The influence of egg survivorship Houde, A.E. and Torio, A.J. (1992) Effect of parasitic on the subsequent nest fidelity of female bicolor infection on male color pattern and female choice in damselfish, Stegastes partitus. Animal Behaviour 46, guppies. Behavioral Ecology 3, 346–51. 111–21. Höglund, J. and Alatalo, R.V. (1995) Leks. Princeton, NJ: Knapp, R.A. and Kovach, J.T. (1991) Courtship as an hon- Princeton University Press. est indicator of male parental quality in the bicolor Jamieson, I.G. (1995) Do female fish prefer to spawn in damselfish, Stegastes partitus. Behavioral Ecology 2, nests with eggs for reasons of mate choice copying or 295–300. egg survival? American Naturalist 145, 824–32. Kodric-Brown, A. (1983) Determinants of male repro- Jennions, M.D. (1997) Female promiscuity and genetic ductive success in pupfish (Cyprinodon pecosensis). incompatibility. Trends in Ecology and Evolution 12, Animal Behaviour 31, 128–37. 251–3. Kodric-Brown, A. (1992) Male dominance can enhance Behavioural Ecology of Reproduction 245

mating success in guppies. Animal Behaviour 44, of parental care and egg cannibalism in the river bull- 165–7. head, Cottus gobio L. (Pisces, Cottidae). Behavioral Kraak, S.B.M. and van den Berghe, E.P. (1992) Do female Ecology and Sociobiology 32, 229–37. fish assess paternal quality by means of test eggs? Marconato, A., Shapiro, D.Y., Petersen, C.W., Warner, Animal Behaviour 43, 865–7. R.R. and Yoshikawa, T. (1997) Methodological analysis Kvarnemo, C. (1994) Temperature differentially affects of fertilization rate in the bluehead wrasse, Thalasso- male and female reproductive rates in the sand ma bifasciatium: pair vs. group spawns. Marine Ecol- goby: consequences for operational sex ratio. Proceed- ogy Progress Series 161, 61–70. ings of the Royal Society of London Series B 256, Maynard Smith, J. (1977) Parental investment: a prospec- 151–6. tive analysis. Animal Behaviour 25, 1–9. Kvarnemo, C. (1998) Temperature modulates competi- Maynard Smith, J. and Harper, D.G.C. (1988) The evolu- tive behaviour: why sand goby males fight more in tion of aggression: can selection generate variability? warmer water. Ethology, Ecology and Evolution 10, Philosophical Transactions of the Royal Society of 105–14. London Series B 319, 557–70. Kvarnemo, C. and Ahnesjö, I. (1996) The dynamics of op- McGovern, J.C., Wyanski, D.M., Pashuk, O., Manooch, erational sex ratios and competition for mates. Trends C.S. II and Sedberry, G.R. (1998) Changes in the sex in Ecology and Evolution 11, 404–8. ratio and size at maturity of gag, Mycteroperca mi- Kvarnemo, C., Svensson, O. and Forsgren, E. (1998) crolepis, from the Atlantic coast of the southeastern Parental behaviour in relation to food availability in United States during 1976–1995. Fisheries Bulletin 96, the common goby. Animal Behaviour 56, 1285–90. 797–807. Ladich, F. (1998) Sound characteristics and outcome McKaye, K.R., Louda, S.M. and Stauffer, J.R.J. (1990) of contests in male croaking gouramis (Teleostei). Bower size and male reproductive success in a cichlid Ethology 104, 517–29. fish lek. American Naturalist 135, 597–613. Leonard, J.L. (1993) Sexual conflict in simultaneous Milinski, M. and Bakker T.C.M. (1990) Female stickle- hermaphrodites: evidence from serranid fishes. En- backs use male coloration in mate choice and hence vironmental Biology of Fishes 36, 135–48. avoid parasitized males. Nature 344, 330–3. Limberger, D. (1983) Pairs and harems in a cichlid fish, Milinski, M. and Bakker, T.C.M. (1992) Costs influence Lamprologus brichardi. Zeitschrift für Tierpsycholo- sequential mate choice in sticklebacks, Gasterosteus gie 62, 115–44. aculeatus. Proceedings of the Royal Society of London Lindström, K. (1992) The effect of resource holding Series B 250, 229–33. potential, nest size and information about resource Mills, S.C. and Reynolds, J.D. (2002) Host preferences quality on the outcome of intruder–owner conflicts in by bitterling (Rhodeus sericeus) spawning in fresh- the sand goby. Behavioral Ecology and Sociobiology water mussels and consequences for offspring sur- 30, 53–8. vival. Animal Behaviour (in press). Lindström, K. and Seppä, T. (1996) The environmental Mousseau, T.A. and Collins, N.C. (1987) Polygyny and potential for polygyny and sexual selection in the nest site abundance in the slimy sculpin (Cottus sand goby. Proceedings of the Royal Society of London cognatus). Canadian Journal of Zoology 65, 2827–9. Series B 263, 1319–23. Munday, P.L., Caley, M.J. and Jones, G.P. (1998) Bi- Lively, C.M., Craddock, C. and Vrijenhoek, R.C. (1990) directional sex change in a coral-dwelling goby. Red Queen hypothesis supported by parasitism in Behavioral Ecology and Sociobiology 43, 371–7. sexual and clonal fish. Nature 344, 864–6. Nakashima, Y., Kuwamura, T. and Yogo, Y. (1996) Both- Lubinski, B.A. Davis, W.P., Taylor, D.S. and Turner, B.J. ways sex change in monogamous coral gobies, (1995) Outcrossing in a natural population of a self- Gobiodon spp. Environmental Biology of Fishes 46, fertilizing hermaphroditic fish. Journal of Heredity 281–8. 86, 469–73. Nilsson, S.Ö. and Nilsson, G.E. (2000) Free choice by fe- Luttbeg, B. (1996) A comparative Bayes tactic for mate as- male sticklebacks: lack of preference for male domi- sessment and choice. Behavioral Ecology 7, 451–60. nance traits. Canadian Journal of Zoology 78, 1251–8. Macías Garcia, C., Jimenez, G. and Contreras, B. (1994) Nordeide, J.T. and Folstad, I. (2000) Is cod lekking or Correlational evidence of a sexually-selected a promiscuous group spawner? Fish and Fisheries 1, handicap. Behavioral Ecology and Sociobiology 35, 90–3. 253–9. Oliveira, R.F., McGregor, P.K. and Latruffe, C. (1998) Macías Garcia, C., Saborío, E. and Berea, C. (1998) Does Know thine enemy: fighting fish gather information male-biased predation lead to male scarcity in vivipa- from observing conspecific interactions. Proceedings rous fish? Journal of Fish Biology 53, 104–17. of the Royal Society of London Series B 265, 1045–9. Marconato, A., Bisazza, A. and Fabris N. (1993) The cost Olson, V.A. and Owens, I.P.F. (1998) Costly sexual 246 Chapter 10

signals: are carotenoids rare, risky or required? Trends Reynolds, J.D. and Gross, M.R. (1992) Female mate pref- in Ecology and Evolution 13, 510–14. erence enhances offspring growth and reproduction in Östlund-Nilsson, S. (2001) Fifteen-spined stickleback a fish, Poecilia reticulata. Proceedings of the Royal So- (Spinachia spinachia) females prefer males with more ciety of London Series B 250, 57–62. secretional threads in their nests: an honest condition Reynolds, J.D. and Jennings, S. (2000) The role of display by males. Behavioral Ecology and Sociobiol- animal behaviour in marine conservation. In: L.M. ogy 50, 263–9. Gosling and W.J. Sutherland (eds) Behaviour and Owens, I.P.F. and Thompson, D.B.A. (1994) Sex differ- Conservation. Cambridge: Cambridge University ences, sex ratios and sex roles. Proceedings of the Press, pp. 238–57. Royal Society of London Series B 258, 93–9. Reynolds, J.D. and Jones, J.C. (1999) Female preference Petersen, C.W. (1987) Reproductive behaviour and gen- for preferred males is reversed under low oxygen con- der allocation in Serranus fasciatus, a hermaphroditic ditions in the common goby (Pomatoschistus mi- reef fish. Animal Behaviour 35, 1601–14. crops). Behavioral Ecology 10, 149–54. Petersen, C.W. (1990) The relationships among popula- Reynolds, J.D., Goodwin, N.B. and Freckleton, R.P. tion density, individual size, mating tactics, and repro- (2002) Evolutionary transitions in parental care and ductive success in a hermaphroditic fish, Serranus live bearing in vertebrates. Philosophical Transactions fasciatus. Behaviour 113, 57–80. of the Royal Society of London B 357, 269–81. Petersen, C.W. and Levitan, D.R. (2001) The Allee effect: Ribbink, A.J. (1987) African lakes and their fishes: con- a barrier to recovery by exploited species. In: J.D. servation scenarios and suggestions. Environmental Reynolds, G.M. Mace, K.H. Redford and J.G. Robinson Biology of Fishes 19, 3–26. (eds) Conservation of Exploited Species. Cambridge: Rogers, W. (1987) Sex ratio, monogamy and breeding Cambridge University Press, pp. 281–300. success in the Midas cichlid (Cichlasoma citrinel- Petersen, C.W. and Marchetti, K. (1989) Filial cannibal- lum). Behavioral Ecology and Sociobiology 21, ism in the cortez damselfish Stegastes rectifraenum. 47–51. Evolution 43, 158–68. Rohwer, S. (1978) Parental cannibalism of offspring and Petersen, C.W. and Warner, R.R. (2002) The ecological egg raiding as a courtship strategy. American Natural- context of reproductive behavior. In: P.F. Sale (ed.) ist 112, 429–40. Coral Reef Fishes. Dynamics and Diversity in a Com- Rosenqvist, G. (1990) Male mate choice and female– plex Ecosystem. San Diego, CA: Academic Press, pp. female competition for mates in the pipefish Nerophis 103–18. ophidion. Animal Behaviour 39, 1110–15. Petersson, E. and Järvi, T. (1997) Reproductive behaviour Ross, R.M. (1990) The evolution of sex-change mecha- of sea trout (Salmo trutta): the consequences of sea- nisms in fishes. Environmental Biology of Fishes 29, ranching. Behaviour 134, 1–22. 81–93. Pickering, H. and Whitmarsh, D. (1997) Artificial reefs Sadovy, Y. (1993) The Nassau grouper, endangered or just and fisheries exploitation: a review of the ‘attraction unlucky? Reef Encounter 13, 10–12. versus production’ debate, the influence of design Sadovy, Y. (1994) Grouper stocks of the western central and its significance for policy. Fisheries Research 31, Atlantic: the need for management and management 39–59. needs. Proceedings of the Gulf Caribbean Fisheries In- Pitcher, T.J. and Parrish, J.K. (1993) Functions of shoaling stitute 43, 43–65. behaviour in teleosts. In: T.J. Pitcher (ed.) Behaviour of Sandvik, M., Rosenqvist, G. and Berglund, A. (2000) Male Teleost Fishes, 2nd edn. London: Chapman & Hall, pp. and female mate choice affects offspring quality in a 363–439. sex-role-reversed pipefish. Proceedings of the Royal Poulin, R. (1994) Mate choice decisions by parasitized Society of London Series B 267, 2151–5. female upland bullies, Gobiomorphus breviceps. Sargent, R.C. (1997) Parental care. In: J.G. Godin (ed.) Be- Proceedings of the Royal Society of London Series B havioural Ecology of Teleost Fishes. Oxford: Oxford 256, 183–7. University Press, pp. 297–315. Qvarnström, A. and Forsgren, E. (1998) Should females Sargent R.C. and Gross, M.R. (1993) William’s principle: prefer dominant males? Trends in Ecology and Evolu- an explanation of parental care in teleost fishes. In: T.J. tion 13, 498–501. Pitcher (ed.) The Behaviour of Teleost Fishes, 2nd edn. Reynolds, J.D. (1996) Animal breeding systems. Trends London: Croom Helm, pp. 333–62. in Ecology and Evolution 11, 68–72. Sato, T. (1986) A brood parasitic catfish of mouthbrood- Reynolds, J.D. and Côté, I.M. (1995) Direct selection on ing cichlid fishes in Lake Tanganyika. Nature 323, mate choice: female blennies pay more for better 58–9. mates. Behavioral Ecology 6, 175–81. Schlupp, I., Marler, C. and Ryan, M.J. (1994) Benefit to Behavioural Ecology of Reproduction 247

male sailfin mollies of mating with heterospecific Vincent, A.C.J., and Sadler, L.M. (1995) Faithful pair females. Science 263, 373–4. bonds in wild seahorses, Hippocampus whitei. Animal Shapiro, D.Y., Marconato, A. and Yoshikawa, T. (1994) Behaviour 50, 1557–69. Sperm economy in a coral reef fish, Thalassoma bifas- Vincent, A.C.J. and Sadovy, Y. (1998) Reproductive ciatum. Ecology 75, 1334–44. ecology in the conservation and management of fishes. Sheldon, B.C. (2000) Differential allocation: tests, mech- In: T. Caro (ed.) Behavioral Ecology and Conservation. anisms and implications. Trends in Ecology and Evo- Oxford: Oxford University Press, pp. 209–45. lution 15, 397–402. Vincent, A.C.J., Ahnesjö, I., Berglund, A. and Rosenqvist, Skarstein, F. and Folstad, I. (1996) Sexual dichromatism G. (1992) Pipefishes and sea-horses: are they all sex role and the immunocompetence handicap: an observa- reversed? Trends in Ecology and Evolution 7, 237–41. tional approach using Arctic charr. Oikos 76, 359– Vrijenhoek, R.C. (1979) Factors affecting clonal diversity 67. and coexistence. American Zoologist 19, 787–97. Smith, C. and Wootton, R.J. (1995) The costs of parental Warner, R.R. (1988a) Traditionality of mating site prefer- care in teleost fishes. Reviews in Fish Biology and ences in a coral reef fish. Nature 335, 719–21. Fisheries 5, 7–22. Warner, R.R. (1988b) Sex change in fishes: hypotheses, St Mary, C.M. (1994) Sex allocation in a simultaneous evidence, and objections. Environmental Biology of hermaphrodite, the blue-banded goby (Lythrypnus Fishes 22, 81–90. dalli): the effects of body size and behavioral gender Warner, R.R. and Hoffman, S.G. (1980) Population den- and the consequences for reproduction. Behavioral sity and the economics of territorial defense in a coral Ecology 5, 304–13. reef fish. Ecology 61, 772–80. Stockley, P., Gage, M.J.G., Parker, G.A. and Møller, A.P. Warner, R.R., Shapiro, D.Y., Marconato, A. and Petersen, (1997) Sperm competition in fishes: the evolution C.W. (1995) Sexual conflict: males with the highest of testis size and ejaculate characteristics. American mating success convey the lowest fertilization bene- Naturalist 149, 933–54. fits to females. Proceedings of the Royal Society of Svensson, I. (1988) Reproductive costs in two sex-role London Series B 262, 135–9. reversed pipefish species (Syngnathidae). Journal of Wedekind, C. (1996) Lek-like spawning behaviour and Animal Ecology 57, 929–42. different female mate preferences in roach (Rutilus Swenson, R.O. (1997) Sex-role reversal in the tide- rutilus). Behaviour 133, 681–95. water goby, Eucyclogobius newberryi. Environmental Wisenden, B.D. (1994) Factors affecting mate desertion Biology of Fishes 50, 27–40. by males in free-ranging convict cichlids (Cichlasoma Taborsky, M. (1984) Broodcare helpers in the cichlid nigrofasciatum). Behavioral Ecology 5, 439–47. fish Lamprologus brichardi: their costs and benefits. Wisenden, B.D. (1999) Alloparental care in fishes. Re- Animal Behaviour 32, 1236–52. views in Fish Biology and Fisheries 9, 45–70. Taborsky, M. (1994) Sneakers, satellites and helpers: Wisenden, B.D. and Keenleyside, M.H.A. (1993) The parasitic and cooperative behaviour in fish reproduc- dilution effect and differential predation following tion. Advances in the Study of Behaviour 23, 1–100. brood adoption in free-ranging convict cichlids Taborsky, M. and Limberger, D. (1981) Helpers in fish. (Cichlasoma nigrofasciatum). Ethology 90, 203–12. Behavioral Ecology and Sociobiology 8, 143–5. Zahavi, A. (1977) Reliability in communication systems Vincent, A.C.J. (1994) Seahorses exhibit conventional and the evolution of altruism. In: B. Stonehouse and sex roles in mating competition, despite male preg- C.M. Perrins (eds) Evolutionary Ecology. London: nancy. Behaviour 128, 135–51. Macmillan, pp. 253–9. Part 3 Fish as Predators and Prey 11 Fish Foraging and Habitat Choice: A Theoretical Perspective

GARY G. MITTELBACH

11.1 INTRODUCTION The diversity of nature presents a challenge to both approaches and there is room for empiricist and Fish come in a variety of shapes and sizes, and their theoretician alike. In writing this chapter, how- diversity in form and the range of environments ever, I have thrown my lot in with those who use that they occupy is unparalleled amongst the models of adaptive behaviour to attempt to pro- vertebrates. Fish as different in size as minnows, vide a general framework for understanding forag- paddlefish (Polyodon spp.), and whale sharks ing behaviours. This field of study is sometimes (Rhincodon typus), make their living feeding referred to as optimal foraging theory, because off tiny zooplankton in the open waters of lakes, optimality models have been used to formulate rivers and oceans. Pike (Esox spp.) and anglerfish hypotheses about adaptive behaviour. A more em- (Lophius spp.) lie in ambush for other fishes, while pirical approach is taken by Juanes et al. (Chapter tuna and swordfish (Xiphias gladius) cruise con- 12, this volume) for piscivores and by Krause et al. tinuously in search of prey. Within a single coral (Chapter 13, this volume) for prey. However, op- reef there is a diversity of feeding modes that timal foraging theory should not be confused with can stagger the imagination. For example, some the notion that natural selection has somehow wrasses (Labridae) make a living cleaning the para- led to the evolution of animals that are ‘optimal’ sites off other fishes, parrotfish (Scaridae) crunch in their behaviours. Rather, as Stephens and corals between their massive jaws and turn them Krebs (1986, p. 213) note, ‘Optimality models . . . into sand, pipefishes (Syngnathidae) and seahorses are yardsticks against which to compare nature; (Hippocampus spp.), with mouths shaped like they are not claims about what nature must be like’. soda-straws, extract tiny invertebrates from their Below, I first discuss the question of diet choice protective crevasses, and triggerfish (Balistidae) in fishes, including the development of optimal blow over sea urchins with a powerful jet of water foraging models, and the potential use of foraging and then feast on their soft undersides. models to predict fish feeding rates and growth in How do we make sense of such a bewildering natural habitats. I then consider problems of habi- array of feeding niches and foraging behaviours? tat selection, using the construct of the ideal free One way would be to classify fish by foraging mode distribution (Fretwell and Lucas 1970; Fretwell or prey type, such as planktivore, molluscivore and 1972) and theories of optimal habitat selection piscivore, and then examine the feeding behaviours based on balancing foraging gains and predation exhibited by fish in each group. In contrast to risks to understand how fish might choose among this empirical approach, one might attempt to de- foraging habitats. Finally, I discuss how habitat velop general theories of foraging behaviour based selection by predator and prey may relate to on evolutionary principles common to all fishes. such important population- and community-level 252 Chapter 11 processes as cascading effects and the response of illustrated in Fig. 11.1 can be modelled using an trophic levels to changes in environmental pro- extension of Holling’s disc equation (Holling ductivity. Mathematical descriptions of theory are 1959a,b) to include attack probabilities and multi- kept to a minimum and I focus instead on provid- ple prey types. In this model, the predator’s feeding ing a conceptual understanding of the theory and rate (net energy gained per foraging time, E/T) is its predictions, and a summary of the main empiri- expressed as: cal tests that have involved fish. Fortunately, fish have played a predominant role in the develop- n ment and testing of adaptive models of foraging ÂliiiiNEP()as P i ()- C s behaviour. i=1 ET= n (11.2) 1+ÂliiiNP() a P i () s H i i=1 11.2 FORAGING BEHAVIOUR AND DIET CHOICE where the net energy value of prey type i, Ei = Aimbi - C H , A is the assimilable fraction of the energy 11.2.1 Feeding preference h i i content of prey type i, bi is the mass of prey type i, Most fish are selective foragers: they prefer to feed m is a conversion constant for prey mass to energy, on some prey types and not on others (Gerking Ch is the energy costs of handling prey, Hi is 1994). Preference can be defined as the difference handling time of prey type i, Cs is the energy cost of between the proportion of a prey type in the diet searching, li is the average per-capita encounter and the proportion of that prey type in the envi- rate with prey of type i, Ni is the average density of ronment. Although feeding preference can be prey type i, Pi(a) is the probability that an encoun- measured by a variety of mathematical indices tered prey of type i is attacked, and Pi(s) is the (Confer and Moore 1987), the index developed by probability that an attacked prey of type i is Manly (1974) and further extended by Chesson successfully consumed. Feeding rate is equal to (1978, 1983) is the most widely accepted. The the total amount of assimilable food that can be Manly–Chesson index calculates the selectivity collected by a fish during one unit of search time, for prey type i as: minus the energy cost of foraging, divided by the total foraging time that results from one unit k of search time (Osenberg and Mittelbach 1989; a = dN dN iii()Â() jj (11.1) Mittelbach and Osenberg 1994). j=1 Preference (positive or negative) for a given prey Where i = 1, 2,..., k and k is the number of prey type may result from factors that influence the rate categories, di is the number (or proportion) of prey at which a prey type is encountered (li) or that of type i in the diet, and Ni is the density (or propor- influence the probability that an encountered tion) of prey of type i in the environment. The prey will be consumed (Pi(a) and Pi(s)). For example, index ai ranges from 0 to 1. Prey types that are larger prey are visible at further distances (Werner consumed in proportion to their abundance in the and Hall 1974; Confer and Blades 1975; Vinyard environment (i.e. no preference) have ai = 1/k; and O’Brien 1976) and prey that are more active are ai > 1/k indicates selection for a prey type and more easily seen by a searching fish (Ware 1973). ai < 1/k indicates selection against a prey type. Any factor that affects the rate of encounter be- Feeding preference may result from active prey tween predator and prey may result in a predator choice by the predator or from a variety of other showing a positive or a negative preference for a factors that may affect the rate of encounter prey type. It is important to note, however, that between predator and prey. We can best under- while differential prey encounter may lead to stand the role of behaviour in fish diet selection by selective foraging, meaning that there is a differ- breaking down the act of predation into its com- ence between the abundance of prey types in the ponent parts (Fig. 11.1). The predation sequence fish’s diet and the abundance of prey types in the Fish Foraging and Habitat Choice 253

Number of prey type i in the environment

Prey size and reactive distance, prey and Probability of predator activity, prey prey encounter λ crypsis, prey refuge use ( iNi)

Number of prey type i encountered

Probability that Active choice by the an encountered predator prey is attacked

(Pi (a))

Number of prey Fig. 11.1 Stages in the predation type i attacked process. The right-hand column lists the components of a multi- species Holling functional Probability that Prey size, mobility and response model (i.e. equation an attacked prey escape behaviour, prey is successfully 11.2) that determine the defence (e.g. shell consumed probability that a given prey type hardness, spines) (P (s)) (i) will be consumed. Listed to the i left are attributes of predator and prey that influence these Number of prey components. (Source: modified type i consumed from Sih 1993.)

environment, this mechanism does not involve 1989; Huckins 1997). The same problem is faced by active predator choice. piscivorous fish. Christensen (1996) and Sih and Predator behaviour may enter into diet selec- Christensen (2001) argue that piscivore diets are tion through a predator’s choice of which prey strongly influenced by the differential capture items to pursue (Pi(a)) and the probability that a success of prey and that this is a major reason why pursuit will result in successful capture (Pi(s)). piscivores generally prefer smaller prey than When prey are small relative to the predator, would be expected based on maximizing their capture success generally will be high. For ex- energy intake per handling time (see Juanes 1994; ample, adult planktivores have very high capture Juanes et al., Chapter 12, this volume). success when feeding on zooplankton. How- ever, even for adult planktivores, some types of 11.2.2 Diet selection and zooplankton such as copepods are sufficiently optimal foraging evasive that fish must use alternative capture techniques (Vinyard 1980). Fish that feed on hard- Much of the theoretical development of foraging bodied prey like molluscs routinely encounter models has revolved around the question of prey that differ in capture success, brought about predator choice: when should a predator choose by a variable probability of successfully crushing a to pursue a prey item that it has encountered? shell (Wainwright 1987; Osenberg and Mittelbach Ecologists and evolutionary biologists have used 254 Chapter 11 optimization criteria to address this question of digestive time may not change the relative rank- predator choice, arguing that natural selection ings of prey types (Kaiser et al. 1992; Nilsson 2000). should result in predator behaviours that maxi- Some studies have shown that prey selection by mize their rate of energy gain, which is a com- a fish changes as the stomach of the fish fills. ponent of fitness. Charnov (1976) developed one A review of this area can be found in Clark and of the first optimal diets models, using an equation Mangel (2000). similar to equation 11.2 to predict a predator’s diet. Fish have also been shown to follow the OFT In Charnov’s model, predators adjust their attack prediction that diets should become more selec- probabilities (Pi(a)) to maximize their total energy tive as the abundance of more profitable prey gain (E/T). Two basic predictions that arise from increases in the environment (Werner and Hall Charnov’s model and from similar models devel- 1974), and in a few cases predicted diets based on oped by Schoener (1971), Pulliam (1974), Werner OFT closely matched observed diets (Mittelbach and Hall (1974) and others are: 1981; Galis and de Jong 1988; Persson and 1 predators should prefer prey that yield more Greenberg 1990). However, in most cases, fish energy per unit handling time; consumed some prey outside the optimal diet. The 2 as the abundance of higher value prey increases reviews by Stephen and Krebs (1986) and Sih and in the environment, lower value prey should Christensen (2001) list 18 studies that tested the be dropped from the diet and predators should predictions of OFT with fish. In only a few of these become more selective (Stephens and Krebs 1986; studies (<20%) was there a good quantitative fit Sih and Christensen 2001). between predicted and observed diets. Given the Fish were used in some of the first tests of opti- simplicity of most optimal diet models, and the mal foraging theory (OFT) (Werner and Hall 1974; fact that these models do not incorporate many Kislalioglu and Gibson 1976; Mittelbach 1981), important components of feeding behaviour such and a number of articles have reviewed the applica- as prey detection and recognition, prey capture tion of OFT to understanding fish diet selection success, predator learning and memory, it is not and feeding behaviour (Werner and Mittelbach surprising that a quantitative fit between theory 1981; Townsend and Winfield 1985; Hart 1986, and observation is missing. However, most (>80%) 1989). In addition, two general reviews of OFT by of the fish studies reviewed by Stephen and Krebs Stephen and Krebs (1986) and Sih and Christensen (1986) and Sih and Christensen (2001) found at (2001) have evaluated the overall success of OFT in least qualitative agreement between predicted and predicting predator diets, including those of fish. observed patterns of diet selection. Further, the What can we conclude from these and other stud- usefulness of OFT may extend beyond the study ies about the applicability of OFT to understand- of feeding behaviour, as OFT provides one of the ing diet choice in fish? As Sih and Christensen few means of predicting how the attack probabil-

(2001) note, this is a bit like asking whether the ities for a prey type (Pi(a) in equation 11.2) should glass is half full or half empty. In many ways, OFT change in different environments. These attack has been successful. Fish do prefer to feed on prey probabilities are needed if we are to use a foraging that yield higher energetic returns. However, we model such as equation 11.2 to predict fish feeding now recognize that other factors besides handling rates under different environmental conditions. time need to be considered in measuring the cost of feeding on a prey item. In particular, for evasive prey, we need to also include the probability of cap- 11.3 FORAGING MODELS ture (Pi(s)) in calculating the expected energetic re- AND FISH GROWTH turn from prey type i (Christensen 1996). Also, for some prey types, digestive time may need to be in- Fish ecologists are often interested in understand- cluded in the total ‘handling’ time for a prey item, ing whether fish will grow better in one lake than although current evidence suggests that including in another or how a change in the prey resource Fish Foraging and Habitat Choice 255 will affect fish growth rates. One approach to ad- (a) 40

dressing this problem is to relate fish feeding rates ) to the abundance of prey in the environment –1 (see also Jobling, Chapter 5, this volume). Foraging models provide a potential link between measures 30 of prey abundance in the field and the expected energy intake of fish. Mittelbach and Osenberg (1994) reviewed four studies that used optimal 20 foraging models to predict the growth or feeding rates of fish in the field. They found that in each case the predictions of the foraging models were 10 0.000 0.200 0.400 well correlated with the performance of the –1 fish. A simpler measure of prey abundance, the Zooplankton biomass (mg l ) total biomass of prey available, was less success- (b) ful at predicting fish growth rates. For example, 40 ) Adult bluegill growth (g year total zooplankton biomass was unrelated to the –1 growth of planktivorous bluegill sunfish (Lepomis macrochirus) in Michigan lakes (Fig. 11.2a), 30 whereas the predicted feeding rate, calculated using equation 11.2 and the available zooplankton size–density distribution, was well correlated 20 with bluegill growth (Fig. 11.2b). Hart and Connellan (1984) also showed that prey profit- ability, expressed as energy gained per handling Adult bluegill growth (g year 10 time, had a significant effect on pike (Esox lucius) 0.0000.100 0.200 0.300 0.400 growth rates in the laboratory. Predicted net energy gain (J s–1) Mittelbach and Osenberg (1994) showed theo- retically that available prey biomass will not Fig. 11.2 (a) Average growth of adult bluegill as a function of the total biomass of zooplankton (dry mass, be a good predictor of fish feeding rates or growth cladocerans only) available in seven Michigan lakes. rates, except under fairly restrictive circum- Each point represents a specific lake/year combination, stances. Therefore, we will generally need to use with each lake represented by a different symbol. some type of foraging model in order to translate (b) Adult bluegill growth as a function of the predicted prey availabilities into expected energy gains to net energy gain calculated using equation 11.2 and the the fish. However, Mittelbach and Osenberg (1994) size–density distribution of zooplankton available in these same lakes. (Source: modified from Mittelbach also found that ‘optimality’ per se was often not a and Osenberg 1994; reproduced by permission of the critical factor in predicting fish foraging gains. University of South Carolina Press.) Rather, it was the effect of prey size on encounter rates and handling times that had the largest impact on predicting fish feeding rates (see also 11.4 FEEDING RATE AND Osenberg and Mittelbach 1989; Persson 1990). GROUP SIZE Thus, the value of using optimal foraging models such as equation 11.2 to translate prey abundances Fish feeding rates may also be affected by group into potential feeding rates may derive more from size. For example, it is often assumed that fish in specifying the critical components of the foraging shoals suffer a decrease in individual feeding rates process, such as encounter rates, than from accu- due to competition (Pitcher 1986). This assump- rately predicting optimal diets. tion may well be true for fish in pelagic habitats (Eggers 1976). However, in structurally complex 256 Chapter 11 habitats foraging in a group may actually increase 11.5.1 Ideal free distribution individual fish feeding rates. For example, Pitcher et al. (1982) demonstrated that goldfish (Carassius In the simplest version of the IFD, food is supplied auratus) and minnows (Phoxinus phoxinus) to a foraging habitat or patch at a fixed rate, where located artificial food clumps in the laboratory it is consumed immediately. If predators are ‘free’ more rapidly when group size increased. However, to select among patches, have perfect or ‘ideal’ Pitcher et al. (1982) did not measure the effect of knowledge of the rates of resource supply and do group size on individual feeding rates, which is the not differ in their competitive abilities, then we ex- critical parameter. Mittelbach (1984) showed that pect the number of predators in a patch to be pro- individual feeding rates of juvenile bluegill hunt- portional to the total food input to that patch. This ing for benthic amphipods first increased with is called the ‘input-matching rule’. A number group size and then decreased. In these experi- of experimental studies have tested the input- ments, foraging rates increased with group size matching prediction, including studies by Milinski because individual fish occasionally ‘flushed’ prey (1979, 1984) with sticklebacks (Gasterosteus that they were unable to capture but were subse- aculeatus), Godin and Keenleyside (1984) with quently captured by other fish in the group. How- cichlids (Aequidens curviceps) and Abrahams ever, data on the effects of group size on individual (1989) with guppies (Poecilia reticulata). These feeding rates in complex environments are very studies and others (see Parker and Sutherland limited. More empirical work is needed to deter- 1986; Tregenza 1995 for reviews) show that the mine if group size affects feeding rate and what number of fish foraging in a patch is often pro- direction this effect may take in different types of portional to the rate of resource input, supporting environments. If feeding rate decreases with group the input-matching prediction of IFD. However, size, while protection from predators increases the further prediction of IFD theory, that the feed- with group size, an optimal group size may result. ing rates of individual foragers should be equal in However, such optimal group sizes are unlikely to all patches, is rarely true (Parker and Sutherland be stable, as individual fish should tend to join a 1986; Hugie and Grand 1998; Tregenza and group rather than remain on their own (Clark and Thompson 1998). A number of factors can lead to Mangel 1984; Pitcher 1986). unequal feeding rates among individuals, includ- ing unequal competitive abilities. Simple IFD mod- els have been modified to include the impact of 11.5 FORAGING AND unequal competitors (e.g. Parker and Sutherland HABITAT SELECTION 1986; Sutherland and Parker 1992), and in these models it is the sum of the competitive units that Ideal free distribution (IFD) theory has guided the is predicted to match the input of resources to a study of habitat selection in fish and other foragers patch. This is called the input-matching of com- since its inception some 30 years ago (Fretwell and petitive units. However, different combinations of Lucas 1970; Fretwell 1972; see also Brown 1969; numbers and types of competitors may satisfy this Orians 1969; Parker 1970). IFD theory is based on prediction. A recent study by Hugie and Grand the optimal foraging principle that animals should (1998) suggests that even when foragers differ in select habitats that maximize their fitness, which their competitive abilities, the expected distribu- is equated with feeding rate in the simplest IFD tion of foragers predicted by IFD models with models, and on the assumption that feeding rates unequal competitors will tend to match the are density dependent. Here I examine the foraging predictions of simple IFD models. Hugie and behaviours of fish in light of IFD theory and general Grand (1998) review a number of studies with fish principles of optimal habitat selection. The litera- that support their hypothesis. ture on this topic is vast and my review only skims Classical IFD models assume a continuous the surface of this conceptually rich subject. input of resources into a patch and immediate Fish Foraging and Habitat Choice 257 consumption of resources by the foragers. These environment, equilibrium may occur. Power’s models differ from ‘standing-stock’ IFD models, (1984) classic study of armoured catfish (Loricari- where food availability in a habitat is characterized dae) feeding on periphyton (attached algae) in a by the density of the resource standing stock Panamanian stream may be one such case. In (Kacelnik et al. 1992; van der Meer and Ens 1997; Power’s (1984) study, algal productivities differed Weber 1998). Natural environments are more among stream pools due to differences in shading. closely mimicked by standing-stock models, The density of armoured catfish was directly pro- where a predator’s feeding rate within a habitat is portional to the percentage of open canopy over the determined by the standing stock of its prey. This, of stream. However, algal standing crops and Lori- course, is also the assumption of classical predator– caridae growth rates, and estimates of their feeding prey models (Taylor 1984). We can use a foraging rates, were similar in both sunny and dark pools. model such as equation 11.2 to predict predator Power’s (1984) results therefore match the pre- feeding rates from prey standing stock, although in dictions of ideal free habitat selection based on most cases a simpler version of Holling’s disc equa- consumer–resource dynamics (Lessells 1995; tion is used. For example, Mittelbach (1981) and Oksanen et al. 1995). Werner et al. (1983a) used a variation of equation A number of studies have considered how inter- 11.2 to predict the potential feeding rates of bluegill ference amongst predators may affect predicted sunfish feeding in the open water, the vegetation habitat distributions in standing-stock IFD mod- and on the bare bottom of Michigan ponds and els (e.g. Sutherland and Parker 1985, 1992; Parker lakes. They found that large bluegill, which were and Sutherland 1986; Ruxton et al. 1992; Lessells not threatened by predators, foraged mostly in the 1995; Moody and Houston 1995; Stillman et al. habitat with the highest predicted energy gain. 1997; van der Meer and Ens 1997). In these models, Thus, the fish behaved according to OFT. predator intake rate is assumed to decrease with The foraging model used by Mittelbach (1981) increasing predator density as the result of inter- and Werner et al. (1983a) assumes that there is no ference. Predicted habitat distributions in these interference between predators and that feeding models are sensitive to the way in which interfer- rates are a function of prey density. In this case, all ence is incorporated into the predator’s functional fish should feed initially in the habitat that pro- response (Stillman et al. 1997; Weber 1998). How- vides the highest foraging return. However, we ever, the results of Free et al. (1977), Lessells (1995) expect relative habitat profitabilities to change as and Oksanen et al. (1995) suggest that patterns prey are depleted within a habitat, causing fish to of habitat use derived from models of ideal free switch habitats. In the studies of Mittelbach (1981) habitat selection and consumer–resource dy- and Werner et al. (1983a), habitat profitabilities did namics may hold true unless interference is quite change through time and large bluegill were able strong. For example, the armoured catfish studied to track these changes, switching habitats to feed by Power (1984) are known to exhibit feeding inter- where there was the highest predicted energy gain. ference, yet their habitat distribution closely Theoretically, if foragers are of equal competitive matched the predictions of IFD theory. ability, this system should settle into an equilib- rium where the distribution of foragers among habitats results in each habitat yielding the same 11.5.2 Incorporating costs feeding rate (Lessells 1995; Oksanen et al. 1995). of habitat use Habitats that have higher prey productivity will Abiotic costs have higher forager densities. However, equilib- rium is unlikely in a strongly seasonal environ- IFD theory assumes that predators pay no cost in ment such as in the bluegill study, where predator choosing habitats, hence the epithet ‘free’. How- and prey dynamics vary in response to changing ever, in nature costs are real. For example, Tyler environmental conditions. In a more constant and Gilliam (1995) considered the abiotic cost of 258 Chapter 11 stream fishes that hold a position and that feed on energy gain, and was their response consistent prey that drift by. They argued that while faster- with maximizing fitness? flowing sites may deliver prey at a higher rate, The difficulty in developing a theory of optimal these sites may be more costly in terms of energy habitat choice that incorporates both predation expended in holding a position and in terms of risk and foraging return is that these factors are a reduced probability of successfully capturing expressed in very different units, predation risk as faster-drifting prey. Both of these costs have been the probability of death per unit time and foraging shown to be important in the choice of habitats by return as energy gained per unit time. Abrahams stream fishes (Fausch 1984; Hughes and Dill 1990; and Dill (1989) approached this problem by using Hill and Grossman 1993). Tyler and Gilliam (1995) a fish’s behaviour to evaluate how much of an in- incorporated these costs into an IFD model and crease in energy gain is required to get a fish to ac- compared the predicted distributions of stream cept a given increase in mortality risk. Abrahams fishes under the IFD with costs to that of the and Dill (1989) used IFD theory to quantify the en- simple IFD, using blacknose dace (Rhinicthys ergetic equivalence of predation risk for guppies atratulus) in a laboratory stream. They found that feeding in two laboratory habitats. In their experi- the IFD incorporating costs did a better job of ment, food was delivered at different rates in the predicting the fishes’ habitat use than did simple two habitats, and they found that in the absence of IFD theory. predators guppies distributed themselves between habitats as predicted by the input-matching rule of IFD theory. Adding a fish predator to one of the Predation risk habitats caused some of the guppies to shift to feed- The natural world is a dangerous place for predator ing in the safer habitat. The difference in per-capita and prey alike (see Krause et al., Chapter 13, this feeding rates between the safe and dangerous habi- volume). Risk of injury or death is another poten- tats therefore provided a measure of the energetic tial cost of habitat choice. Fish and other organ- equivalence of predation risk. This assumed that isms have been shown to respond to predation an individual’s predation risk does not change with risk in choosing habitats. For example, Mittelbach group size (Moody et al. 1996). In a second experi- (1981) and Werner et al. (1983a) found that large ment, Abrahams and Dill (1989) increased feeding bluegills, which were relatively immune to preda- rates in the risky habitat by the amount predicted tors, foraged in the habitats that yielded the to offset the influence of predation risk. They highest energy gain. However, smaller bluegill, found that the additional food resulted in a similar which were vulnerable to predation by largemouth number of guppies using both the risky and safe bass (Micropterus salmoides), were more limited habitats; female guppies, however, fit the pre- in their habitat use. Small bluegills fed in the pro- dicted response better than males. Grand and Dill tection of the littoral zone vegetation, even though (1996) adopted this same approach to examine the open-water habitat yielded a higher feeding the use of cover by stream-dwelling coho salmon rate and better growth (Werner et al. 1983b; see (Oncorhynchus kisutch). In nature, cover provides also Werner and Hall 1988). Thus, small bluegill the salmon with protection from predators but appeared to trade-higher energy gain for a lowered may also result in reduced food availability. Grand predation risk. These and other studies (Milinski and Dill (1996) used IFD theory to quantify the en- and Heller 1978) prompted theoreticians and em- ergetic equivalence of cover to the fish and then piricists to investigate the balance between forag- calculated how much additional food would have ing gain and predation risk, focusing on two main to be added to a risky patch in order to make it of questions: (i) what was the optimal habitat choice equal value to a safe patch. When they added this that would maximize fitness and (ii) how did fish additional amount of food, the fish returned to and other foragers respond to predation risk and the distribution they had before the risk became Fish Foraging and Habitat Choice 259 evident. Thus, the experiments of Abrahams and use of fishes, and in the past 15 years several hun- Dill (1989) and Grand and Dill (1996) show that dred papers have been published on this topic (see fish can balance energy gain and predation risk in reviews in Dill 1987; Lima and Dill 1990; Lima choosing habitats and that IFD theory provides a 1998a,b). While many of these studies show that means of empirically equating these two factors. fish respond to the foraging gain/predation risk The above studies provide an example of an em- trade-off in an adaptive manner, few studies have pirical approach to equating energy gain and preda- actually tested specific models of optimal habitat tion risk in determining habitat choice. However, choice in fish. Gilliam and Fraser (1987) tested a in order to predict habitat use a priori, predation modification of the m/g rule in the habitat choice of risk and foraging gain must be modelled in a com- juvenile creek chubs (Semotilus atromaculatus). mon currency associated with individual fitness. They showed that when fish have the choice be- Gilliam (1982) developed the first of these models, tween feeding in two or more habitats that differ using the methods of optimal control theory. in energy gain and predation risk, and when the Stephens and Krebs (1986) provide a description of fish may also use a refuge that contains no food the theory and see also Werner and Gilliam (1984). and therefore has no predation risk, the m/g rule Gilliam (1982) found that in the simplest case, collapses to the simpler rule, ‘when foraging, mini- where fish are of pre-reproductive size, population mize the ratio of mortality rate to gross foraging rate size is constant and mortality rate within a habitat (f)’. Gilliam and Fraser (1987) tested the ‘minimize is constant, then the habitat choice that maxi- m/f’ prediction by offering juvenile creek chubs a mizes a fish’s fitness is selection of the habitat that choice between two foraging areas that differed in minimizes the ratio of mortality rate (m) to growth experimentally manipulated resource densities rate (g). That is, select the habitat which provides and mortality risk from adult creek chubs. They a unit of growth at the lowest mortality cost. found that the creek chubs choice of foraging habi- This result was quickly labelled the ‘m/g rule’ or tats agreed well with the theoretical predictions Gilliam’s rule and it had a profound effect on the (Fig. 11.3). In this study, the optimal habitat choice development of subsequent theory of habitat se- boiled down to a rather simple rule of thumb: prefer lection incorporating predation risk and foraging habitat A over habitat B if mA/fA predator satiation or predator confu- Do fish select habitats so as to optimally balance sion (Milinski and Heller 1978). A number of foraging gain and predation risk using the ‘mini- researchers have considered how changes in pre- mize m/g rule’ or some variant thereof? Clearly, a dation risk with group size may impact habitat large number of studies show that both predation choice within the context of IFD theory (e.g. risk and foraging gain may influence the habitat McNamara and Houston 1990; Moody et al. 1996); 260 Chapter 11

(a) tive difference in competitive abilities among for- Resource in Predicted 1.0 1-predator switch agers (Moody et al. 1996; Grand and Dill 1999). site The more empirical aspects of predation are discussed by Juanes et al. (Chapter 12, this volume) + and of prey avoidance of risk through refuge by 0.5 Krause et al. (Chapter 13, this volume).

+ 11.5.3 Habitat choice in a 0 three-trophic-level system A forager’s risk of mortality within a habitat (b) Predicted may also change due to the movement of their own 1.0 switch predators between habitats. In a three-trophic- level system, predators at the top trophic level and + + foragers at the middle trophic level are both free to Preference for more hazardous site 0.5 select between habitats that differ in resources at + + the bottom trophic level. In this situation, habitat choice by the predators and foragers can be mod- elled as a spatial game. Hugie and Dill (1994) mod- 0 0.2 0.4 0.6 0.8 1.0 elled the situation where there are two habitats Tubifex density in more hazardous site (cm–2) that differ in their inherent riskiness, indepen- dently of predator densities, and that differ in the Fig. 11.3 Predicted and observed habitat selection by rate of resource production. Foragers and predators juvenile creek chubs: (a) habitat preference in are free to select between habitats and forager feed- experiments where habitats contained either one or three predators; (b) habitat preference in experiments ing rates follow the assumption of IFD theory, with with habitats containing one or two predators. The resource production within a habitat being divided x-axis shows the density of resources (Tubifex worms) equally among foragers. Predator feeding rates are in the habitat with the higher mortality risk; the y-axis based on their functional response to prey density, shows the observed habitat use (measured as preference which also determines the prey’s mortality rate, for the riskier habitat). The vertical arrows labelled and predators, being at the top trophic level, are ‘Predicted switch’ indicate the density of resources in assumed to have a constant mortality rate among the riskier habitat, where the fish are predicted to switch from feeding in the safer habitat to feeding in the riskier habitats. Both predator and prey are assumed to habitat. Points represent the results of individual feeding choose habitats so as to maximize their net repro- trials. Crosses represent treatment means; horizontal ductive rate (R0), which is an appropriate measure bars represent pooled means for orthogonal contrasts. of fitness only when population size is constant (Source: modified from Gilliam and Fraser 1987; over time (Stearns 1992; see Hutchings, Chaper reproduced by permission of the Ecological Society of 7, this volume for alternative fitness measures). America.) Maximizing R0 is the same fitness criterion used by Gilliam (1982), and Hugie and Dill’s (1994) cri- Grand and Dill (1999) recently modelled this terion of optimal habitat choice by both predator case for foragers of equal and unequal competitive and prey is equivalent to Gilliam and Fraser’s ability. In general, these models predict that the criterion of minimizing m/f. dilution of risk with group size results in habitat Hugie and Dill (1994) found that when habitats distributions that differ from the IFD input- differ in their inherent riskiness, independently matching rule. Specific predictions, however, of predator density, foragers should prefer the less depend on the relative risk of using alternative risky habitat, and the ratio of forager densities in habitats, the strength of risk dilution and the rela- the two habitats should be the inverse of the ratio Fish Foraging and Habitat Choice 261

No interference (a) (b) 2.5 14 12 2.0 10 ' j j 1.5

d 8 d / '/ i i d d 1.0 6 2.5 4 2.5 0.5 2.0 2 2.0 1.5 1.5 1.0 Pi/Pj 0 1.0 0.0 0.5 0.5 0.5 0.5 Pi/Pj 1.0 1.5 0.0 1.0 1.5 0.0 2.0 2.5 2.0 2.5 Ri/Rj Ri/Rj

Interference (c) (d) 2.5 14 2.0 12 10

j 1.5 ' d j / d 8 i d '/ 1.0 i d 6 2.5 0.5 2.0 4 2.5 1.5 2 2.0 1.0 P /Pj 1.5 0.0 i 0.5 0.5 0 1.0 1.0 1.5 0.0 0.5 Pi/Pj 2.0 2.5 0.5 1.0 1.5 2.0 0.0 R /Rj 2.5 i Ri/Rj

Fig. 11.4 Results of the habitat selection game between predator and prey. Habitats i and j differ in inherent riskiness

(R) and productivity of the prey’s food resource (P). All panels show the ratio of prey (di/dj) and predator (di¢/dj¢) densities as a function of the ratios of habitat riskiness (Ri/Rj) and resource productivity (Pi/Pj). (a,b) Results for the basic model (no interference between predators); (c,d) results when predators interfere with each other’s foraging ability. Note that in (a) and (b), predator density responds to differences in habitat productivities but prey density does not. In (c) and (d), both predator and prey densities respond to habitat productivities. (Source: modified from Hugie and Dill 1994.) of the habitat risks (Fig. 11.4a). For example, if one of foragers (Fig. 11.4a). Predator distributions, habitat is twice as risky as another, it should have however, are dependent on both relative habitat half the density of prey. There is of course consider- riskiness and relative habitat productivity, and able empirical support for the prediction that fish more predators will be found in the more produc- and other foragers should prefer the inherently less tive habitats (Fig. 11.4b). risky habitat (see reviews in Lima and Dill 1990; The results of Hugie and Dill’s (1994) behav- Lima 1998a,b). The more surprising prediction of ioural game between predator and prey closely Hugie and Dill’s (1994) model is that habitat pro- parallel the predictions of ecological food web ductivity should have no effect on the distribution theory (Hairston et al. 1960; Oksanen et al. 1981) 262 Chapter 11 based on trophic-level responses to changes in en- not achieve a simultaneous IFD. Schwinning and vironmental productivity. In simple food-chain Rosenzweig (1990) present some of the factors that models with three trophic levels, the biomasses of might stabilize such oscillations. the top and bottom trophic levels are predicted to Adaptive responses by fish to predation risk and respond positively to an increase in the productiv- foraging gain may also lead to cascading effects ity of the environment, whereas the biomass of the in aquatic food webs. A number of studies have middle trophic level remains constant (Oksanen shown that a change in the abundance of piscivo- et al. 1981). In these models, changes in biomass rous fish can have effects that cascade down the among trophic levels are due to numerical res- food chain to lower trophic levels, although these ponses brought about through births and deaths may be buffered by redundancy in marine food and not to movement of individuals between webs (see Polunin and Pinnegar, Chapter 14 and habitats of differing productivity (but see Wootton Persson, Chapter 15, this volume; and Kaiser and and Power 1993). The results of Hugie and Dill’s Jennings, Chapter 16, Volume 2). Such top-down (1994) analysis show that the same pattern of effects are typically assumed to be transmitted response to an increase in productivity within a by direct consumption of lower trophic levels by habitat (i.e. an increase in the abundance of the top higher trophic levels. However, behavioural re- trophic level and no change in the abundance of the sponses by fish and other foragers may produce the middle trophic level) can result from behavioural same type of trophic cascade, if an increase in the decisions of predators and foragers, with no abundance of a higher trophic level causes foragers change in global population size by the consumers. at the next lower trophic level to reduce feeding ac- Wootton and Power (1993) suggest the same result, tivity, shift habitats or spend more time in a refuge although their theoretical development of a behav- (e.g. Turner and Mittelbach 1990; Brabrand and ioural model is limited to a short appendix to their Faafeng 1993). paper. Hugie and Dill (1994) also note that if pre- dators exhibit interference, then the abundance of both predators and foragers will depend on habitat 11.6 CONCLUSIONS productivity (Fig. 11.4c,d), which again is analo- gous to the findings of food-chain population mod- In this chapter, I have used models of adaptive be- els with interference among the top carnivores haviour to provide a framework for understanding (Wollkind 1976; Mittelbach et al. 1988). the foraging behaviour of fish. IFD theory, OFT and Hugie and Dill (1994) conclude their paper with ‘m/g rules’ all attempt to predict an animal’s forag- the statement that ‘Perhaps the single most impor- ing behaviour based on the premise that energy tant new insight arising from our modeling of acquisition is central to the fitness of an organism. habitat selection as a game is the prediction that The power of the optimality approach is that it pro- prey distributions may not respond to experimental vides a mechanism to predict, a priori, how a fish changes in food abundance, whenever their preda- chooses its diet or selects its habitat in a given tors are free to respond in a dynamic way’. I would environment. These predictions can then be com- suggest that an equally interesting and exciting pared to the forager’s observed behaviour in the result is that behavioural models and population laboratory or in the field. It is at this point, where dynamic models predict the same pattern of model predictions are compared with empirical trophic-level response to environmental produc- data, that a major bone of contention arises con- tivity, albeit for different reasons. Schwinning cerning the use of optimality models. Observed and Rosenzweig (1990) examined a model system diets or patterns of habitat selection never pre- in which all three trophic levels were free to choose cisely match the predictions of theory (see reviews among habitats based on IFD criteria. They found by Stephens and Krebs 1986; Sih and Christensen that in general the system oscillated through time, 2001). To some, this implies that we should reject with individuals shifting back and forth between the theory and move on (e.g. Pierce and Ollason habitats such that the three trophic levels could 1987). To others, qualitative agreement between Fish Foraging and Habitat Choice 263 theory and data suggests that the general approach for sharing their thoughts and insights. The is valid and that a closer look at the mismatch US National Science Foundation has generously between theory and data may provide a means to supported our work on fish foraging over the years, develop better models (e.g. Krebs and McCleery for which I am grateful. This is KBS contribution 1984; Stephens and Krebs 1986). While much has # 901. been written on this debate and other issues con- cerning the use of optimality models in biology, the topic remains contentious (Perry and Pianka REFERENCES 1997; Sih and Christensen 2001). It is clear that quantitative studies measuring Abrahams, M.V. (1989) Foraging guppies and the ideal energy gain and mortality risk, or other costs of free distribution: the influence of information on patch choice. Ethology 82, 116–26. foraging, have provided a powerful framework for Abrahams, M.V. and Dill, L.M. (1989) A determination understanding the foraging behaviours of fish. No of the energetic equivalence of the risk of predation. doubt, future studies of foraging theory will con- Ecology 70, 999–1007. tinue to wrestle with questions of how fish and Abrams, P.A. (1990) The effects of adaptive-behavior on other animals may evaluate potential energy gain, the type-2 functional-response. Ecology 71, 877–85. predation risk and other foraging costs in the ‘real Abrams, P.A. (1991) Strengths of indirect effects gener- ated by optimal foraging. Oikos 62, 167–76. world’ where foragers are less than ‘ideal’ in their Brabrand, Å. and Faafeng, B. (1993) Habitat shift in roach knowledge and where behavioural choices are (Rutilus rutilus) induced by pikeperch (Stizostedion never ‘free’. However, while the continued devel- lucioperca) introduction: predation risk versus pelagic opment and testing of foraging theory is impor- behaviour. Oecologia 95, 38–46. tant, I suggest that the opportunity to use foraging Brown, J.L. (1969) The buffer effect and productivity in tit models as tools to study the population and populations. American Naturalist 103, 347–54. community ecology of fishes is equally valuable. Charnov, E.L. (1976) Optimal foraging: attack strategy of a mantid. American Naturalist 110, 141–51. Incorporating adaptive foraging behaviour into Chesson, J. (1978) Measuring preference in selective predator–prey models can provide new insights predation. Ecology 59, 211–15. into the predator functional response and pre- Chesson, J. (1983) The estimation and analysis of prefer- dator–prey dynamics (e.g. Abrams 1990, 1991). ence and its relationship to foraging models. Ecology Optimal foraging models may also provide a link 64, 1297–304. between measures of prey abundance in the field Christensen, B. (1996) Predator foraging capabilities and and the predicted feeding rates of fish (Persson prey antipredator behaviours: pre- versus postcapture constraints on size-dependent predator–prey interac- 1990; Mittelbach and Osenberg 1994). Finally, tions. Oikos 76, 368–80. habitat selection behaviour provides an alter- Clark, C. and Mangel, M. (1984) Foraging and flocking native mechanism for generating important strategies: information in an uncertain environment. population- and community-level phenomena American Naturalist 123, 626–41. in aquatic ecosystems, such as cascading effects Clark, C.W. and Mangel, M. (2000) Dynamic State Vari- and trophic-level responses to productivity. able Models in Ecology. Methods and Applications. Oxford: Oxford University Press. Confer, J.L. and Blades, P.I. (1975) Omnivorous zooplank- ton and planktivorous fish. Limnology and Oceanog- ACKNOWLEDGEMENTS raphy 20, 571–9. Confer, J.L. and Moore, M.V. (1987) Interpreting selectiv- I thank Nathan Dorn, Christopher Steiner and ity indexes calculated from field data or conditions of Jeremy Wojdak for comments on an earlier draft of prey replacement. Canadian Journal of Fisheries and the manuscript. Numerous discussions with col- Aquatic Sciences 44, 1529–33. Dill, L.M. (1987) Animal decision making and its ecologi- leagues and friends have shaped my understanding cal consequences: the future of aquatic ecology and of fish foraging behaviours, and I especially thank behaviour. Canadian Journal of Zoology 65, 803–11. James Gilliam, Craig Osenberg, Peter Wainwright, Eggers, D.M. (1976) Theoretical effect of schooling Earl Werner and my former and current students by planktivorous fish predators on rate of prey 264 Chapter 11

consumption. Journal of the Fisheries Research Board Holling, C.S. (1959a) The components of predation as of Canada 33, 1965–71. revealed by a study of small mammal predation on Fausch, K.D. (1984) Profitable stream positions for the European pine sawfly. Canadian Entomologist salmonids: relating specific growth rate to net energy 91, 293–320. gain. Canadian Journal of Zoology 62, 441–51. Holling, C.S. (1959b) Some characteristics of simple Free, C.A., Beddington, J.R. and Lawton, J.H. (1977) On types of predation and parasitism. Canadian Entomol- the inadequacy of simple models of mutual interfer- ogist 91, 385–98. ence for parasitism and predation. Journal of Animal Houston, A.I. and McNamara, J.M. (1999) Models of Ecology 46, 543–54. Adaptive Behaviour. Cambridge: Cambridge Univer- Fretwell, S.D. (1972) Populations in a Seasonal Environ- sity Press. ment. Princeton, NJ: Princeton University Press. Houston, A.I., McNamara, J.M. and Hutchinson, J.M.C. Fretwell, S.D. and Lucas, H.L. (1970) On territorial (1993) General results concerning the trade-off behavior and other factors influencing habitat dis- between gaining energy and avoiding predation. Philo- tribution in birds. I. Theoretical development. Acta sophical Transactions of the Royal Society of London Biotheoretica 19, 16–36. Series B 341, 375–97. Galis, F. and de Jong, P.W. (1988) Optimal foraging and Huckins, C.J.F. (1997) Linkages between morphology, ontogeny: food selection by Haplochromis piceatus. feeding performance, diet, and competitive ability of Oecologia 75, 175–84. molluscivorous sunfish. Ecology 78, 2401–14. Gerking, S.D. (1994) Feeding Ecology of Fish. New York: Hughes, N.F. and Dill, L.M. (1990) Position choice by Academic Press. drift-feeding salmonids: model and test for Arctic Gilliam, J.F. (1982) Foraging under mortality risk in size- grayling (Thymallus arcticus) in subarctic mountain structured populations. PhD thesis, Michigan State streams, interior Alaska. Canadian Journal of Fish- University, East Lansing, MI. eries and Aquatic Sciences 47, 2039–48. Gilliam, J.F. and Fraser, D.F. (1987) Habitat selection when Hugie, D.M. and Dill, L.M. (1994) Fish and game: a game foraging under predation hazard: a model and a test with theoretic approach to habitat selection by predators stream-dwelling minnows. Ecology 68, 1856–62. and prey. Journal of Fish Biology 45 (Suppl. A), 151– Godin, J.-G. and Keenleyside, M. (1984) Foraging on 69. patchily distributed prey by a cichlid fish (Teleostei, Hugie, D.M. and Grand, T.C. (1998) Movement between Cichlidae): a test of the ideal free distribution theory. patches, unequal competitors, and the ideal free distri- Animal Behaviour 32, 120–31. bution. Evolutionary Ecology 12, 1–19. Grand, T.C. and Dill, L.M. (1996) The energetic equiva- Juanes, F. (1994) What determines prey size selectivity in lence of cover to juvenile coho salmon (Oncorhynchus piscivorous fishes. In: D.J. Stouder, K.L. Fresh and R.J. kisutch): ideal free distribution theory applied. Behav- Feller (eds) Theory and Application in Studies of Fish ioral Ecology 8, 437–47. Feeding Ecology. Columbia, SC: University of South Grand, T.C. and Dill, L.M. (1999) Predation risk, unequal Carolina Press, pp. 79–100. competitors and the ideal free distribution. Evolution- Kacelnik, A., Krebs, J.R. and Bernstein, C. (1992) The ary Ecology Research 1, 389–409. ideal free distribution and predator–prey populations. Hairston, N.G., Smith, F.E. and Slobodkin, L.B. (1960) Trends in Ecology and Evolution 7, 50–5. Community structure, population control, and com- Kaiser, M.J., Westhead, A.P., Hughes, R.N. and Gibson, petition. American Naturalist 94, 421–5. R.N. (1992) Are digestive characteristics important Hart, P.J.B. (1986) Foraging in teleost fishes. In: T.J. contributors to the profitability of prey? Oecologia 90, Pitcher (ed.) The Behaviour of Teleost Fishes. 61–9. Baltimore, MD: Johns Hopkins University Press, pp. Kislalioglu, M. and Gibson, R.N. (1976) Some factors gov- 211–35. erning prey selection by the 15-spined stickleback, Hart, P.J.B. (1989) Predicting resource utilization: the Spinachia spinachia (L.). Journal of Experimental utility of optimally foraging models. Journal of Fish Marine Biology and Ecology 25, 159–69. Biology 35 (Suppl. A), 271–7. Krebs, J.R. and McCleery, R.H. (1984) Optimization Hart, P.J.B. and Connellan, B. (1984) The cost of prey cap- in behavioural ecology. In: J.R. Krebs and N.B. ture, growth rate and ration size in pike, Esox lucius L., Davies (eds) Behavioural Ecology: An Evolutionary as functions of prey weight. Journal of Fish Biology 25, Approach, 2nd edn. Oxford: Blackwell Scientific, 279–92. pp. 91–121. Hill, J. and Grossman, G.D. (1993) An energetic model of Lessells, C.M. (1995) Putting resource dynamics into microhabitat use for rainbow trout and rosyside dace. continuous input ideal free distribution models. Ecology 74, 685–98. Animal Behaviour 49, 487–94. Fish Foraging and Habitat Choice 265

Lima, S.L. (1998a) Nonlethal effects in the ecology of Nilsson, A. (2000) Pikeivory: behavioural mechanisms predator–prey interactions. BioScience 48, 25–34. in northern pike piscivory. PhD dissertation, Univer- Lima, S.L. (1998b) Stress and decision-making under sity of Lund, Lund, Sweden. the risk of predation: recent developments from Oksanen, L., Fretwell, S.D., Arruda, J. and Niemela, P. behavioral, reproductive, and ecological perspectives. (1981) Exploitation ecosystems in gradients of pri- Advances in the Study of Behavior 27, 215–90. mary productivity. American Naturalist 118, 240– Lima, S.L. and Dill, L.M. (1990) Behavioral decisions 61. made under the risk of predation: a review and pro- Oksanen, T., Power, M.E. and Oksanen, L. (1995) Ideal spectus. Canadian Journal of Zoology 68, 619–40. free habitat selection and consumer–resource dynam- Ludwig, D. and Rowe, L. (1990) Life-history strategies ics. American Naturalist 146, 565–85. for energy gain and predator avoidance under time Orians, G.H. (1969) On the evolution of mating systems constraints. American Naturalist 135, 686–707. in birds and mammals. American Naturalist 203, Mangel, M. and Clark, C.W. (1986) Towards a unified 589–603. foraging theory. Ecology 67, 1127–38. Osenberg, C.W. and Mittelbach, G.G. (1989) Effects of Mangel, M. and Clark, C.W. (1988) Dynamic Modeling body size on the predator–prey interaction between in Behavioural Ecology. Princeton, NJ: Princeton pumpkinseed sunfish and gastropods. Ecological University Press. Monographs 59, 405–32. Manly, B.F.J. (1974) A model for certain types of selection Parker, G.A. (1970) The reproductive behaviour and experiments. Biometrics 30, 281–94. temporal relationships of each sex around the site of McNamara, J.M. and Houston, A.I. (1990) State- mating and oviposition. Journal of Animal Ecology dependent ideal free distributions. Evolutionary 39, 205–28. Ecology 4, 298–311. Parker, G.A. and Sutherland, W.J. (1986) Ideal free dis- Milinski, M. (1979) An evolutionary stable feeding strat- tributions when individuals differ in competitive egy in sticklebacks. Zeitschrift für Tierpsychologie 51, ability: phenotype-limited ideal free models. Animal 36–40. Behaviour 34, 1222–42. Milinski, M. (1984) Competitive resource sharing: an Perry, G. and Pianka, E.R. (1997) Animal foraging: past, experimental test of a learning rule for ESS’s. Animal present and future. Trends in Ecology and Evolution Behaviour 31, 233–42. 12, 360–4. Milinski, M. and Heller, R. (1978) The influence of Persson, L. (1990) Predicting ontogenetic niche shifts in a predator on the optimal foraging behaviour of the field: what can be gained by foraging theory? In: stickleback (Gasterosteus aculeatus). Nature 275, R.N. Hughes (ed.) Behavioural Mechanisms of Food 642–4. Selection. Berlin: Springer-Verlag, pp, 303–21. Mittelbach, G.G. (1981) Foraging efficiency and body Persson, L. and Greenberg, L.A. (1990) Optimal foraging size: a study of optimal diet and habitat use by and habitat shift in perch (Perca fluviatilis) in a bluegills. Ecology 62, 1370–86. resource gradient. Ecology 71, 1699–713. Mittelbach, G.G. (1984) Group size and feeding rates in Pierce, G.J. and Ollason, J.G. (1987) Eight reasons why bluegills. Copeia 1984, 998–1000. optimal foraging theory is a complete waste of Mittelbach, G.G. and Osenberg, C.W. (1994) Using time. Oikos 49, 111–17. foraging theory to study trophic interactions. In: D.J. Pitcher, T.J. (1986) Functions of shoaling behaviour in Stouder, K.L. Fresh and R.J. Feller (eds) Theory teleosts. In: T.J. Pitcher (ed.) The Behavior of Teleost and Application in Studies of Fish Feeding Ecology. Fishes. Baltimore, MD: Johns Hopkins University Columbia, SC: University of South Carolina Press, Press, pp. 294–337. pp. 151–70. Pitcher, T.J., Magurran, A.E. and Winfield, I.J. (1982) Fish Mittelbach, G.G., Osenberg, C.W. and Leibold, M.A. in larger shoals find food faster. Behavioral Ecology (1988) Trophic relations and ontogenetic niche shifts and Sociobiology 10, 149–51. in aquatic ecosystems. In: B. Ebenman and L. Persson Power, M.E. (1984) Habitat quality and the distribution (eds) Size-structured Populations: Ecology and Evolu- of algae-grazing catfish in a Panamanian stream. tion. Berlin: Springer-Verlag, pp. 219–35. Journal of Animal Ecology 53, 357–74. Moody, A.L. and Houston, A.I. (1995) Interference Pulliam, H.R. (1974) On the theory of optimal diets. and the ideal free distribution. Animal Behaviour American Naturalist 108, 59–75. 49, 1065–72. Rowe, L. and Ludwig, D. (1991) Size and timing of meta- Moody, A.L., Houston, A.I. and McNamara, J.M. (1996) morphosis in complex life cycles: time constraints and Ideal free distributions under predation risk. Behav- variation. Ecology 72, 413–27. ioral Ecology and Sociobiology 38, 131–43. Ruxton, G.D., Gurney, W.S.C. and de Roos, A.M. (1992) 266 Chapter 11

Interference and generation cycles. Theoretical Popu- Tyler, J.A. and Gilliam, J.F. (1995) Ideal free distributions lation Biology 42, 235–53. of stream fish: a model and test with minnows, Rhinic- Schoerner, T.W. (1971) Theory of feeding strategies. thys atratulus. Ecology 76, 580–92. Annual Review of Ecology and Systematics 2, 369– van der Meer, J. and Ens, B. (1997) Models of interference 404. and their consequences for the spatial distribution of Schwinning, S. and Rosenzweig, M.L. (1990) Periodic ideal and free predators. Journal of Animal Ecology 66, oscillations in an ideal-free predator–prey distribu- 846–58. tion. Oikos 59, 85–91. Vinyard, G.L. (1980) Differential prey vulnerability and Sih, A. (1993) Effects of ecological interactions on forager predator selectivity: effects of evasive prey on bluegill diets: competition, predation risk, parasitism and prey (Lepomis macrochirus) and pumpkinseed (L. gibbosus) behaviour. In: R.N. Hughes (ed.) Diet Selection: an predation. Canadian Journal of Fisheries and Aquatic Interdisciplinary Approach to Foraging Behaviour. Sciences 37, 2294–9. Oxford: Blackwell Scientific Publications, pp. 182– Vinyard, G.L. and O’Brien, W.J. (1976) Effects of light and 211. turbidity on the reactive distance of bluegill sunfish Sih, A. and Christensen, B. (2001) Optimal diet theory: (Lepomis macrochirus). Journal of the Fisheries when does it work and when and why does it fail? Research Board of Canada 33, 2845–9. Animal Behaviour 61, 379–90. Wainwright, P.C. (1987) Morphology and ecology: the Sih, A., Ziemba, R. and Harding, K.C. (2000) New in- functional basis of feeding constraints in Caribbean sights on how temporal variation in predation risk labrid fishes. Ecology 69, 635–44. shapes prey behavior. Trends in Ecology and Evolution Ware, D.M. (1973) Risk of epibenthic prey to predation by 15, 3–4. rainbow trout (Salmo gairdneri). Canadian Journal of Stearns, S.C. (1992) The Evolution of Life Histories. Fisheries and Aquatic Sciences 30, 787–97. Oxford: Oxford University Press. Weber, T.P. (1998) News from the realm of the ideal Stephens, D.W. and Krebs, J.R. (1986) Foraging Theory. free distribution. Trends in Ecology and Evolution Princeton, NJ: Princeton University Press. 13, 89–90. Stillman, R.A., Goss-Custard, J.D. and Caldow, R.W.G. Werner, E.E. and Gilliam, J.F. (1984) The ontogenetic (1997) Modelling interference from basic foraging niche and species interactions in size-structured popu- behaviour. Journal of Animal Ecology 66, 692–703. lations. Annual Review of Ecology and Systematics Sutherland, W.J. and Parker, G.A. (1985) Distribution of 15, 393–425. unequal competitors. In: R.M. Sibly and Smith, R.H. Werner, E.E. and Hall, D.J. (1974) Optimal foraging (eds) Behavioural Ecology. Ecological Consequences and the size selection of prey by the bluegill sunfish of Adaptive Behaviour. Oxford: Blackwell Scientific (Lepomis macrochirus). Ecology 55, 1042–52. Publications, pp. 255–73. Werner, E.E. and Hall, D.J. (1988) Ontogenetic habitat Sutherland, W.J. and Parker, G.A. (1992) The relationship shifts in bluegill: the foraging rate–predation risk between continuous input and interference models trade-off. Ecology 69, 1352–66. of ideal free distributions with unequal competitors. Werner, E.E. and Mittelbach, G.G. (1981) Optimal forag- Animal Behaviour 44, 345–56. ing: field tests of diet choice and habitat switching. Taylor, R.J. (1984) Predation. London: Chapman & Hall. American Zoologist 21, 813–29. Townsend, C.R. and Winfield, I.J. (1985) The application Werner, E.E., Mittelbach, G.G., Hall, D.J. and Gilliam, of optimal foraging theory to feeding behaviour in fish. J.F. (1983a) Experimental tests of optimal habitat use in In: P. Tytler and P. Calow (eds) Fish Energetics: New fish: the role of relative habitat profitability. Ecology Perspectives. Baltimore, MD: Johns Hopkins Univer- 64, 1525–39. sity Press, pp. 67–98. Werner, E.E., Gilliam, J.F., Hall, D.J. and Mittelbach, Tregenza, T. (1995) Building on the ideal free distribu- G.G. (1983b) An experimental test of the effects of pre- tion. Advances in Ecological Research 26, 253–307. dation risk on habitat use in fish. Ecology 64, 1540–8. Tregenza, T. and Thompson, D.J. (1998) Unequal com- Wollkind, D.J. (1976) Exploitation in three trophic levels: petitor ideal free distribution in fish? Evolutionary an extension allowing intraspecies carnivore interac- Ecology 12, 655–66. tion. American Naturalist 110, 431–47. Turner, A.M. and Mittelbach, G.G. (1990) Predator Wootton, J.T. and Power, M.E. (1993) Productivity, avoidance and community structure: interactions consumers, and the structure of a river food chain. among piscivores, planktivores, and plankton. Ecology Proceedings of the National Academy of Sciences 71, 2241–54. USA 90, 1384–7. 12 Feeding Ecology of Piscivorous Fishes

FRANCIS JUANES, JEFFREY A. BUCKEL AND FREDERICK S. SCHARF

12.1 INTRODUCTION group, however, are difficult to categorize and describe, which perhaps explains why no reviews Fish exhibit tremendous diversity in feeding of their ecology exist. habits and the morphologies associated with feeding. A recent book (Gerking 1994) and various other overviews of fish feeding exist (Wootton 12.2 ADAPTATIONS FOR 1990; Hobson 1991; Hart 1993). However, most of PISCIVORY these tend to be general reviews of theory or focus 12.2.1 What is a piscivore? on smaller non-piscivorous fishes. Our intent here is to review the feeding ecology of piscivorous fish, We define piscivorous fish as carnivorous fish that a subject which to our knowledge has never previ- consume primarily fish prey. Most fish species are ously been reviewed. We focus on teleost fish, and opportunistic and flexible in their feeding habits on species and sizes that consume juvenile and (Dill 1983) and no species consumes only fish prey; adult prey and do not consider those that feed pri- however many do ingest fish as the main prey marily on larvae or fish eggs. item. Fish that eat other fish are second in propor- tion to those feeding on benthic invertebrates 12.1.1 Why piscivorous fish? and are present in a variety of freshwater, estuarine and marine systems. They are equally common in Piscivorous fish are broadly distributed phyloge- tropical and temperate ecosystems. Keast (1985) netically and geographically, occur in most habi- examined the piscivore feeding guild of small lakes tats and generally occupy the top of most aquatic and streams and categorized piscivorous fish into trophic webs. Piscivorous fish also generally primary and secondary piscivores. Primary or achieve the largest body size within fish com- ‘specialized’ piscivores shift to piscivory within munities, are represented by some of the largest the first few months of life, whereas secondary pis- species (many elasmobranchs, tunas, billfishes) civores only become fish-eaters later in life. Keast and have potentially large impacts on their com- further suggests that secondary piscivores switch munities through predation. Finally, many pisci- to fish as a way to maintain energetic efficiency vorous fishes, because of their ubiquity and large as they grow. This can only be achieved by eating body size are among the most valuable harvestable progressively larger prey and, at a certain point, species in many of the world’s fisheries. In some fish are the only prey available. Furthermore, sec- cases, they compete with humans for commer- ondary piscivores are not structurally adapted for cially important resource species (V. Christensen piscivory other than acquiring a large mouth as 1996; Buckel et al. 1999a). Piscivorous fish as a they age. 268 Chapter 12

12.2.2 What are the adaptations strikes occur at high speed and missed prey are for piscivory? rarely pursued, or they can be pursuers such as trout (Salmo trutta), where attacks start from a C-shaped Feeding behaviours and position, strikes occur at low speeds from a short morphological adaptations distance and missed prey are chased. When observed in detail it becomes clear that pis- civore feeding behaviour is complex and flexible in Ambush This strategy is used by species that dealing with different prey types. Behaviours can attack from seclusion, although an element of generally be grouped into the following categories, chasing may also occur. Examples are morays which are also associated with particular morpho- (Muraenidae), pike (Esocidae) and summer floun- logical adaptations. der (Pleuronectidae). Few morphological similar- ities exist in species that use this behaviour other Luring Luring is a sit-and-wait behaviour where than the ability to use camouflage and high-speed prey are attracted by a ‘lure’, which consists of a attacks. However, distinct patterns of kinematic, stalk topped by a device that resembles a source of pressure, electromyographic and behavioural pro- food; the lure is a modification of the first dorsal files of prey capture are exhibited by ambushers spine (Gerking 1994). Luring is typical of the an- and pursuers. glerfish (order Lophiiformes). In addition, among the frogfish various morphological features allow Other A variety of other rarer feeding habits each the predator to blend into the rocky environments with their own specialized morphologies, particu- in which they live as a form of camouflage. Angler- larly in the dentition, include forms of parasitism fish have a considerable gape and ingest prey sizes such as blood-sucking in lampreys (Petromyzonti- that are large compared to other piscivores. dae) and catfishes (Trichomycteridae and Cetopsi- dae), scale-eating in many cichlids and characoids, Stalking Stalking is the unobtrusive pursuit of and fin and eye-biting (Gerking 1994). prey before the attack occurs. This strategy is com- A novel feeding behaviour common to plankti- mon to trumpetfish (Aulostomidae), longnose gar vores has recently been observed for a few species (Lepisosteidae) and needlefish (Belonidae). These of piscivores (Sazima 1998). Ram suspension (or species have similar morphologies, long slender filter) feeding is defined as swimming through the bodies with a long snout and sharp teeth. water with the mouth wide open and opercles flared as a way of filtering small prey items out of Chasing Large piscivores are able to chase and the water column without directing attacks to- ‘swim-down’ prey. This strategy is best represented wards individual prey. This may be a mechanism by the billfish (Xiphiidae and Istiophoridae) and used by piscivores feeding on relatively small prey some tuna (Scombridae) and yellowtail or pompano in high concentrations (B. Hanrahan and F. Juanes, (Carangidae). These fish must be able to attain personal observation). high cruising and accelerating speeds. Their bodies can be described as thunniform or carangiform, Most piscivores ingest their prey whole. A few where thrust is maximized by a lunate tail with a species are able to tear off and ingest pieces and high aspect ratio, a narrow caudal peduncle that good examples are piranhas (Serrasalmus spp.) minimizes sideways thrust and a large anterior and African tiger fish (Hydrocyon vittatus). The body depth that minimizes recoil of the head end. most detailed analysis of partial prey eating has These same features, along with a relatively rigid been performed for bluefish (Pomatomus and streamlined body, also minimize drag (see saltatrix) (Juanes and Conover 1994a; Scharf et al. Brix, Chapter 4, this volume). Chasers can be either 1997). Juvenile bluefish switch from taking prey lungers, such as the pike (Esox spp.), where attacks whole to partial prey consumption when the prey are started at close range from an S-shaped position, to predator size ratio is about 0.35 independent of Feeding Ecology of Piscivorous Fishes 269 prey type (Scharf et al. 1997). This ability allows length, has a fineness ratio between 3.5 and 5.0 bluefish to attack much larger prey sizes than can (Juanes et al. 1994). Interestingly, offshore inverte- predators of similar size and is likely to be a func- brate-feeding juveniles (<40mm) are already mor- tion of specialized musculature and dentition. phologically specialized for piscivory, suggesting Bluefish are also unique in that they ingest prey a trade-off between feeding efficiency and future tail-first, whereas most other piscivores that have diet. A result of this trade-off may be to accelerate been examined eat prey head-first. Tail-first inges- the onset of piscivory. tion may be a result of pursuing prey rather than ambushing it, or a way to reduce prey mobility Life history thereby increasing prey vulnerability, as has been observed in piranhas. Onset of piscivory Most piscivorous fish under- go ontogenetic shifts in diet (Werner and Gilliam 1984; Keast 1985; Winemiller 1989). These shifts Schooling in predators generally progress from consumption of zooplank- Schooling by prey is generally thought of as an ton to consumption of benthic macrofauna or prey adaptation for evading, confusing and reducing the fish, with a concomitant increase in mean prey efficiency of predators (Pitcher and Parrish 1993). size as predators grow. There is much variation in Schooling also has hydrodynamic and foraging the timing of the shift to piscivory among primary functions. Little empirical work has been done on and secondary piscivores (Mittelbach and Persson whether schooling by predators enhances preda- 1998; Mittelbach, Chapter 11, this volume). A tion efficiency in piscivores. Field studies have few species of scombrids apparently forgo the shown that, in general, group attacks tend to result zooplanktivorous stage and start eating fishes at in higher capture success rates (Pitcher and Parrish first feeding, whereas others shift early in the 1993). Eklöv (1992) has shown that the foraging larval period (Tanaka et al. 1996). The shift to efficiency, measured as growth rate, varies be- piscivory invariably results in an increase in pre- tween a group-foraging, actively searching pisci- dator growth rate (Buijse and Houthuijzen 1992; vore (Eurasian perch, Perca fluviatilis) and a Juanes and Conover 1994b; Olson 1996). Among solitary-foraging, stalking piscivore (pike, Esox the scombrids studied by Tanaka et al. (1996), there lucius). Under similar conditions, grouped perch was a direct correlation between the age at the grew more than single perch or grouped pike. In onset of piscivory and early growth, with those contrast, solitary pike grew better than solitary species shifting to piscivory at first feeding capable perch or grouped pike. Schooling in piscivores may of reaching 100mm during the first month of life. have foraging costs and benefits. Prey encounter Within fish cohorts, the largest individuals are rates may be maximized but the probability of los- often able to switch to piscivory while the smallest ing a prey item to kleptoparasitism is also greater are delayed and experience reduced growth. This when predators attack in groups (Juanes and effect leads to the often-observed bimodality in Conover 1994a). There has been anecdotal infor- size distribution of juvenile piscivores (Adams and mation collected that suggests that some piscivo- DeAngelis 1987; Frankiewicz et al. 1996). Because rous fish hunt cooperatively (Pitcher and Parrish survivorship is generally size-selective in fish 1993), although potential mechanisms have not (Sogard 1997), bimodality can result in increased been quantified under controlled conditions. mortality of individuals in the smaller size mode Active piscivorous pelagic fishes are character- (Olson 1996). ized by a range of length to maximum depth ratios (the fineness ratio), which ranges from 4.0 to 6.5 What allows species to shift to piscivory? The and which maximizes feeding efficiency and mini- timing of the onset of piscivory depends on predator mizes drag (Blake 1983). Bluefish, a primary pisci- morphology, predator and prey phenologies, prey vore that switches to fish prey at about 40mm total abundance and abiotic factors. Clearly predators 270 Chapter 12 have to be larger than their prey in order to be able and walleye (Forney 1971; Keast 1985). Among to ingest them. Large body size and large mouth marine fish it is much more difficult to examine gapes are therefore generally considered an impor- the timing of predator and prey spawning. How- tant constraint on prey use, particularly piscivory ever, bluefish appear to have evolved a life-history (Werner 1977; Wainwright and Richard 1995; strategy whereby they are spawned at approxi- Mittelbach and Persson 1998). Biomechanical fea- mately the same temperature as their future prey tures such as jaw mechanics and tooth develop- but at a more southern latitude. This divergence ment have also been implicated (Jenkins et al. 1984; allows bluefish to attain the size advantage re- Wainwright and Richard 1995). Among piscivo- quired when they shift to piscivory after being rous scombrids, ontogenetic development of the advected to higher latitude estuaries where their digestive system (Tanaka et al. 1996) and develop- prey are abundant. ment of the visual system (Margulies 1997) have In a recent review of the freshwater fish litera- been correlated with the onset of piscivory. Wall- ture, Mittelbach and Persson (1998) showed that eye (Stizostedion vitreum) generally shift to pis- those species born larger and with larger mouth civory between 30 and 60mm in length but are able gapes become piscivorous at younger ages and to ingest larvae when they are as small as 10mm; smaller sizes, although they did not find a relation- the earlier onset of piscivory may be related to ship between spawning temperature, used as an in- the abundance of prey in appropriate size ranges dicator of timing, and size at the shift to piscivory (Mathias and Li 1982). However, because the or size at age 1. However, species that accelerated switch to piscivory is most often preceded by an the onset of piscivory by shifting early were larger invertebrate-feeding stage, piscivores must ensure at age 1 and continued to be larger through later rapid growth during that phase so that when fish ages. prey are available the feeding shift can occur. If A potential cost of early onset of piscivory is growth is slowed during the invertebrate-feeding reduced growth and survival if habitat condi- stage, the piscivore size advantage over its fish prey tions dictate prey abundance. Furthermore, be- is reduced leading to delays in the shift to piscivory cause most piscivores are strongly size-selective (Olson 1996). Environmental factors such as low (Juanes 1994), if availability of appropriately sized temperatures early in the growing season can fish prey is delayed, growth may be reduced also reduce growth and delay piscivory (Buijse and (Buckel et al. 1998). Houthuijzen 1992; Olson 1996). Resource polymorphisms Is acceleration of the onset of piscivory adaptive? Because of the dramatic increase in growth rate Adaptive trophic specialization can in some cases following the shift to piscivory and because size- lead to the evolution of resource polymorphisms, selective mortality is so prevalent among juvenile where trophically and morphologically special- fishes (Sogard 1997), natural selection should ized morphs can coexist and ultimately speciation favour life-history strategies in piscivores that can occur. Among fishes various examples of minimize the length of the zooplanktivorous trophic polymorphisms include piscivorous phase. One such strategy to attain an acceleration morphs. For example, in Arctic charr (Salvelinus of the onset of piscivory would be to match the alpinus), four trophic morphs are often recognized. timing and location of spawning with the avail- The four morphs have differing life-history strate- ability of fish prey. Freshwater primary piscivores gies, morphologies and behaviours that are gene- have evolved the ability to spawn earlier than tically based (Skulason et al. 1993). Using other fishes, thereby attaining sufficient size to life-history theory, Mangel (1996) has shown that allow them to consume juvenile fish spawned later the evolution of a large piscivorous morph in trout in the same year. This pattern has been observed (‘ferox’ trout) can occur if the growth rates of in largemouth bass (Micropterus salmoides), pike the different asymptotic morph sizes differ and if Feeding Ecology of Piscivorous Fishes 271 there is size-dependent mortality. Other aspects of on prey activity producing encounters and the pos- alternative life-history evolution caused by com- itive relationship that exists between body size, petitive reproductive behaviour are reviewed by prey movement and detectability. Hutchings (Chapter 7, this volume). The estimation of encounter rates between fish predators and prey remains an elusive problem in fish ecology. Specifically, for piscivorous fishes, 12.3 COMPONENTS OF the mobility of potential prey means that behav- PREDATION iours and movement patterns of both predator and 12.3.1 Search, encounter, pursuit, prey will ultimately affect rates of encounter. Due capture, handling to the variable nature of the plethora of environ- mental parameters that can influence the distribu- For a prey fish to be included in the diet of a pre- tion of animals in space and time, encounter rates dator, it has to be located, pursued, captured, in situ are virtually impossible to measure with manipulated or handled, and finally digested any certainty. Laboratory tanks used for piscivores (Mittelbach, Chapter 11, this volume). Each of and their prey are generally too small to yield any these steps must be performed successfully by the realistic estimates of encounter rate, as all prey are predator in order for the predation process to be usually within the visual field of a given predator. complete. Therefore, prey have several opportun- Therefore, encounter probabilities are often esti- ities to avoid being eaten during the course of the mated using mathematical models that incorpo- interaction. The first of these is to not be detected. rate average swimming velocities of predator Predators have evolved various modes of searching and prey as well as a predator detection radius for prey. At the same time, prey have evolved an (Gerritsen and Strickler 1977). Although theore- assortment of behaviours to reduce their chances tical models can be quite elegant, they remain bur- of being detected by potential predators (Krause dened with many assumptions that often cannot et al., Chapter 13, this volume). The probability be tested. Because no predation can occur without of any one prey being encountered will then an encounter first, much future research will be be determined by the morphological and behavi- devoted to better understanding the factors that oural characteristics of both predator and contribute to variable rates of encounter. prey. Once prey are encountered, predators must be Most predators that consume motile prey, such sufficiently capable of capturing prey for it to be as fish, utilize one of two basic search/encounter eaten. Piscivorous fishes generally have lower modes: ambush or sit-and-wait tactics and cruis- capture probabilities compared with planktivores. ing tactics as described earlier. The types and sizes Usually about half of piscivore attacks on average of prey eaten by a predator are often associated are successful compared with 70–80% for plank- with the foraging tactics employed. For example, tivorous fishes. Capture success in piscivores has Greene (1986) showed that invertebrate predators also been shown to be directly related to relative that employed ambush tactics consumed larger body sizes of prey and predator, and generally de- prey relative to cruising predators. He further clines linearly as the prey size to predator size ratio demonstrated that prey encounters for ambush increases (Fig. 12.1). Similar size-based capture predators were strongly dependent on prey swim- success functions have been observed for several ming speeds, whereas variation in prey activity piscivore species representing different life stages had no discernible effect on encounter rates for (Miller et al. 1988; Juanes and Conover 1994a). cruising predators. Among fish, ambush predators Differences in prey type have also been shown to have also been shown to take larger prey relative to influence predator capture success, as prey escape more actively searching predators. The inclusion proficiencies vary among species (Wahl and Stein of larger prey items in the diets of ambush preda- 1988; Scharf et al. 1998). The strong dependence of tors is thought to be a function of their dependence piscivore capture success on prey size and type 272 Chapter 12

Capture success Handling Profitability time

Predation component Fig. 12.1 General empirical relationships between predation components and relative prey Prey size/predator size size for piscivores.

indicates that it plays an important role in deter- prey type, between encounter probability and mining relative vulnerabilities to predation for prey size will determine whether the profitability different prey. peak is shifted toward smaller or larger relative Similar to capture success, handling time in pis- prey sizes. Accurate profitability functions that civorous fishes is affected by both prey size and incorporate a broad range of prey and predator sizes type. Because larger prey sizes often need to be ma- can be extremely valuable for predicting piscivore nipulated before swallowing, increasing prey size diets. typically causes an exponential increase in han- dling time (Fig. 12.1). The rate of increase in han- 12.3.2 Functional and dling time has been shown to be unique for specific prey types (Hoyle and Keast 1987; Scharf et al. numerical responses 1998). Handling time has historically been a criti- Predatory response to variations in prey density cal parameter to estimate, as it has frequently been can be grouped into two categories. The functional used to represent the primary energetic cost of response refers to the number of prey eaten per feeding in theoretical foraging models attempting predator per unit time as a function of prey density. to predict predator diet (Mittelbach, Chapter 11, Changes in number of predators in response to this volume). Although foraging models have variations in prey density describes the numerical demonstrated some success in predicting diets of response. planktivorous fishes, past models for piscivores have often failed (Sih and Christensen 2001). More Functional responses recent models that incorporate differential capture probabilities based on prey size and type have There are several different forms of functional re- proven more successful in predicting piscivore sponse. A subset of these have been classified as diets (Rice et al. 1993). The combination of size- the type I, II and III models (Holling 1965). The type dependent encounter rates, capture probabilities, I response occurs when the number of prey eaten energy content and handling times can be used to per predator per unit time increases linearly with construct profitability functions for specific prey increasing prey density (Fig. 12.2a). The proportion types. For piscivores, these functions are typically of prey density consumed stays constant and this dome-shaped curves that peak at intermediate response is also called density-independent preda- ratios of prey size to predator size (Fig. 12.1). tion (Fig. 12.2d). Most vertebrates have a type II or However, most often encounter rates cannot be III functional response (Hassell 1978). The type II estimated with confidence and curves are con- response is represented by a negatively accelerat- structed using only capture success and handling ing curve (Fig. 12.2b); therefore, the proportion of time functions. The specific relation, for each prey density consumed decreases with increasing Feeding Ecology of Piscivorous Fishes 273

(a) (d) Type I (Ruggerone and Rogers 1983) and southern floun- der (Paralichthys lethostigma) preying on spot (Leiostomus xanthurus) (Wright et al. 1993). Lake trout (Salvelinus nemaycush) were capable of high predation rates at low prey fish densities, suggest- Type I ing a type II response (Eby et al. 1995). Petersen and DeAngelis (1992) found that type II and III models (b) (e) Type II gave similar fits to field data of northern squawfish (Ptychocheilus oregonensis) feeding on juvenile salmonids. Although the type III functional response has not been observed with regularity, it has been Type II Proportion consumed proposed as a potential mechanism in regulating population abundance (see Section 12.6). Several (c) (f) Type III factors can lead to a type III functional response, such as predator learning, prey refuge and the pres- ence of alternative prey (Holling 1965). The pres- Number of prey consumed per predator unit time ence and accessibility of a prey refuge has been hypothesized to be a factor leading to positive den- Type III sity-dependent mortality in some piscivore–prey Number of prey present systems (Bailey 1994; Hixon and Carr 1997). Alter- native prey can lead to a type III functional re- Fig. 12.2 (a–c) Shapes of hypothetical functional sponse through switching behaviour (Murdoch responses: (a) type I, (b) type II and (c) type III. (d–f) and Oaten 1975). Prey switching occurs when the Proportion of prey consumed or percentage mortality prey type with the highest relative abundance is corresponding to the functional responses: (d) type I, included disproportionately more in the predator’s (e) type II and (f) type III. diet than would be expected from random feeding (see Section 12.4.1 for an example). prey density and is referred to as negative density- Functional responses are critical parts of the dependent predation. Under this regime the risk of multispecies virtual population analysis (MSVPA) being preyed upon is high at low prey densities but method of modelling fish populations (Shepherd decreases with increasing prey density (Fig. 12.2e). and Pope, Chapter 7, Volume 2). MSVPA differs Positive density-dependent predation can result from single-species virtual population analysis when a predator feeds with a type III functional re- (SSVPA) in that interspecific and intraspecific pre- sponse. The shape of this function is sigmoidal dation are included to estimate levels of natural (Fig. 12.2c); when the proportion of prey density mortality, which in SSVPA is assumed constant. consumed is plotted the slope is initially positive, Generally, MSVPA has assumed a type II fun- meaning that the risk of being preyed upon is small ctional response, which may be reasonable for at low prey densities but increases up to a certain predators that do not depend on a small number of point as prey density increases (Fig. 12.2f). Preda- prey species, as in temperate systems, but may not tion components such as attack rate and handling be adequate for predators that do depend on few time (Section 12.3.1) can be estimated from these prey species, as in boreal systems. Inclusion of a functions (Holling 1965; Hassell 1978). type III response can have important effects on In general, piscivorous fishes have a type II forecasts based on MSVPA, including producing functional response. For example, type II responses more than one solution to MSVPA equations were found in Arctic charr preying on migrating (Magnússon 1995; see also Shepherd and Pope, sockeye salmon (Oncorhynchus nerka) smolts Chapter 7, Volume 2). 274 Chapter 12

Numerical responses experiments and large-scale field manipulations to determine patterns of prey selection in several Predator numbers can respond rapidly to increas- freshwater predators, including muskellunge ing prey density, as when predators aggregate on (Esox masquinongy), tiger muskellunge, and prey fish, or can respond slowly, such as increased northern pike. Results demonstrated that each of reproductive success (Peterman and Gatto 1978). the predators consistently selected for shad prey as We know little about numerical responses in pis- opposed to Centrarchid prey. The authors conclud- civorous fish. This is surprising given its potential ed that behavioural and morphological differences importance in regulating population levels; the between prey types resulted in differential vulner- lack of data stems from the logistical difficulties ability to predation and the observed patterns of of measuring such a response. Whenever a predator selection. Research on the feeding habits of increases in density in response to increasing prey piscivorous bluefish in marine waters off the fish density, the total predation rate response can northern Atlantic coast of the USA have also re- differ from the predator’s functional response. A vealed evidence for selective predation on specific sigmoidal response (type III) may more typically prey types. Buckel et al. (1999b) demonstrated describe a piscivore’s total response; the total that, during early summer in shallow estuarine response is a function of both the functional and habitats, juvenile bluefish consumed young-of- numerical response (Bailey 1994). the-year striped bass (Morone saxatilis) in greater proportions than a random sampling of the prey environment, and that this positive selection was 12.4 PREY TYPE AND directly related to striped bass abundance. In other SIZE SELECTIVITY words, they showed prey switching behaviour as referred to in Section 12.3. During southerly au- Selective predation can be defined as the consump- tumn migrations in continental shelf waters, juve- tion of prey in different proportions than those nile bluefish have also shown the propensity for available in the predator’s surrounding habitat selecting specific prey types, particularly bay (Chesson 1978; Mittelbach, Chapter 11, this vol- anchovy (Anchoa mitchilli) (Buckel et al. 1999c). ume). This is in contrast to random feeding, where Selection for specific prey sizes may be even more predators feed indiscriminately on prey in accord- prevalent than prey type selection among piscivo- ance with their relative availability. Predators rous fishes. A recent review of 32 laboratory and that feed in a random manner with respect to rela- field studies concluded that selection for small tive prey abundance levels are usually referred to prey from available size distributions was a com- as opportunistic feeders because they adjust their mon phenomenon for both freshwater and marine feeding habits rapidly to match variation in the piscivores (Juanes 1994). This study also revealed local prey field. Predators that consume a narrow that most piscivorous fishes consumed prey small- range of prey types or sizes regardless of changes er than the optimal prey sizes predicted by energy in the prey field are thought of as specialists and maximization models, independent of both represent the most extreme case of selective predator and prey type as well as predator size predation. Most predators fall somewhere along a (Mittelbach, Chapter 11, this volume). continuum between opportunists and specialists. 12.4.2 Behavioural mechanisms of 12.4.1 Observed patterns selective feeding Numerous instances of prey type and size selec- Selective feeding can result from both morphologi- tion by piscivorous fishes have been reported in cal constraints of predators and behavioural inter- both freshwater and marine communities. For actions between piscivores and their prey. Predator example, Wahl and Stein (1988) used laboratory gape size ultimately limits the sizes of prey in- Feeding Ecology of Piscivorous Fishes 275 gested and can also affect the types of prey con- alter attack and capture rates of predators and con- sumed. However, recent studies indicate that be- tribute significantly to the observed patterns of havioural capabilities of both predator and prey selective feeding. can regulate the sizes and types of prey eaten before morphological constraints become important. In a well-designed series of laboratory experiments, B. 12.5 PREDATOR-SIZE AND Christensen (1996) evaluated the effects of prey an- PREY-SIZE RELATIONSHIPS tipredator behaviours on the sizes of roach (Rutilus rutilus) consumed by piscivorous Eurasian perch. Because they are the top predators in many aquatic Feeding experiments were conducted at two dif- systems, knowledge of the sizes of prey included in ferent spatial scales, determined by tank volumes, the diets of piscivorous fishes is essential in order to to control the level of antipredator behaviour ex- identify their potential impact on prey survival and pressed by the prey. In smaller arenas that limited their role in structuring populations at lower troph- prey escape ability, the author found that perch ic levels. This is particularly important for the were able to consume large roach approaching gape ecosystem approach to fisheries management, limitations. In larger arenas, where prey antipreda- where knowledge of interactions is critical (Pauly tor behaviours were not suppressed, the maximum and Christensen, Chapter 10, Volume 2). Further size of roach consumed by perch was significantly discussion of how size-dependent processes struc- smaller. The author concluded that the sizes of ture communities can be found in Persson, Chapter prey eaten were largely dependent upon relative 15, this volume. From a behavioural standpoint, predator and prey mobility (Mittelbach, Chapter relative body sizes of prey and predator can have 11, this volume). Juanes (1994) hypothesized that significant effects on predator feeding success. common patterns of selection for small-sized prey Detection and capture of prey by piscivores are by piscivorous fishes were the result of size- enhanced with increasing size for several reasons, dependent vulnerabilities of prey to predator including increased sustained and burst swimming capture, rather than predator preferences. He speeds and better visual acuity (Webb 1976). The proposed that many examples of selective feeding escape proficiency of prey also varies with ontogeny in fishes were actually passive selective processes as reaction distances increase and swimming abili- rather than active predator choice, with attack ties are improved with size (Brix, Chapter 4, this rates being relatively equal among different prey volume). Morphological characteristics that influ- sizes but smaller prey being consumed more fre- ence piscivore–prey interactions, such as predator quently due to ease of capture (see also Hart and gape size and robustness of prey morphological Hamrin 1990). Prey behaviour can also influence defences (e.g. spines), also scale with ontogeny. the rate of attack among different prey types Therefore, identifying patterns of prey size use by available to a predator. For example, laboratory ex- piscivorous fishes can provide important clues as periments demonstrated higher attack rates by to the mechanisms that shape piscivore diets and goby (Gobiusculus flavescens) predators on her- the effects of predation by piscivorous fishes on ring (Clupea harengus) larvae compared with cod community structure. (Gadus morhua) larvae, with attack proportions being strongly related to differential levels of prey 12.5.1 General patterns activity between the species (Utne-Palm 2000). and hypotheses Predictions of prey selection from most traditional foraging models assume that optimal choices are For piscivorous fishes, the sizes of prey consumed mainly the result of behavioural decisions by generally increase with predator size. In addition, predators (Stephens and Krebs 1986). For piscivo- the range of prey sizes eaten typically increases in rous fish predators, the behavioural and foraging larger predators, as maximum prey size often in- capabilities of both predator and prey can clearly creases rapidly while minimum prey size may 276 Chapter 12

butions eaten by fish predators of varying body sizes. Based on ratio-scale analyses, he concluded that the range of prey sizes eaten did not change significantly with increasing body size for most predators and that older larger predators should not compete with smaller ones. Although predator size–prey size relationships for piscivorous fishes are often reported, potential mechanisms that lead to the commonly observed Prey size pattern of increasing maximum prey sizes coupled with stationary minimum prey sizes are generally lacking. Scharf et al. (2000) noted that the con- tinued inclusion of small prey in the diets of larger predators contrasts with predictions of optimal foraging models for particulate feeders, which in- dicate that the largest prey available should be con- Predator size sumed preferentially to maximize net energetic return. The authors hypothesized that the combi- Fig. 12.3 Hypothetical predator size–prey size nation of high relative abundance and high capture relationship for a piscivore, illustrating typically probability for small prey relative to large prey may observed changes in minimum and maximum prey sizes consumed with increasing predator size. lead to consistently high vulnerability to preda- tion for small prey fishes as already discussed. Because predator handling times also increase change only slightly over a broad range of predator rapidly with prey size, they suggested that the sizes (Fig. 12.3). For example, Mittelbach and retention of small prey in the diet may reflect Persson (1998) demonstrated increasing mean and profitable foraging decisions by predators because maximum prey sizes as body size increased for 12 search, capture and handling costs remain low. species of freshwater piscivores. The authors further noted that for many of the piscivores they 12.5.2 What determines examined, minimum prey size remained fairly maximum prey size? constant with increasing predator body size. Scharf et al. (2000) revealed similar predator Maximum prey sizes eaten by piscivores can be size–prey size patterns for a group of 18 marine pis- limited by several factors. Of foremost importance civores in continental shelf waters off the Atlantic are morphological limitations imposed by preda- coast of the USA. Expanding prey-size ranges that tor gape size and throat width. Most piscivorous retain small prey in the diet of larger piscivores fishes swallow prey head first after manipulating means that prey sizes consumed by smaller preda- prey so that the largest body depth is positioned tors can be a subset of the prey sizes consumed laterally in the mouth (Reimchen 1991). A strong by larger predators, which may result in a competi- relationship between prey body depth and predator tive disadvantage for smaller predators (Wilson gape width has been detected for several piscivo- 1975). However, when the range of prey sizes eaten rous fishes (e.g. Hambright et al. 1991). General is examined as a ratio of predator size rather than mouth shape and structure can also affect the on an absolute scale, ontogenetic changes are more types and sizes of prey that can be ingested. For subtle. For example, Pearre (1986) studied a large example, Keast and Webb (1966) demonstrated group of predators that included 43 species of strong relationships between mouth morphology larvae, juveniles and some small adult fishes in and the feeding ecology of several freshwater an effort to detect general trends in prey-size distri- fishes. The effects of piscivore gape limitations on Feeding Ecology of Piscivorous Fishes 277 maximum prey sizes eaten can also impact prey (natural) levels of mortality: secondly, that populations. Persson et al. (1996) examined the they do not collapse at all quickly, when potential community-level effects of predator gape subjected to high levels of mortality. limitation in a field study of predation in European lakes. Prey fishes in four lakes were exposed to two One of the more well accepted mechanisms predators with different gape sizes and their popu- put forward to explain population regulation in lations monitored. Results indicated that prey ex- fish is density-dependent mortality in larval and posed to the smaller-gaped predator reached a size juvenile stages through interspecific predation refuge sooner, were more abundant and had slower or cannibalism (see Section 12.6.2; Myers, Chapter growth rates due to intraspecific competition com- 6, this volume; Shepherd and Pope, Chapter 8, pared with prey exposed to the larger-gaped preda- Volume 2). tor. The effects of piscivore gape size were also Density-dependent mortality can lead to the observed at lower trophic levels, indicating the maintenance of population stability (Murdoch and potential importance of predator gape sizes on Oaten 1975; Hassell 1978). Field studies provide community dynamics (Persson, Chapter 15, this evidence for density-dependent mortality in volume). marine fishes (Myers and Cadigan 1993; Forrester Other factors may restrict maximum prey sizes 1995; Hixon and Carr 1997). Several of these inves- eaten by fish predators before morphological limi- tigations provide evidence that piscivorous fishes tations imposed by gape size become important. are the dominant cause of density-dependent mor- The behavioural tactics used by predators to tality. However, identifying the mechanism that search for, encounter and attack prey can affect generates density-dependent predation mortality prey sizes consumed. For example, Gaughan and has been elusive (Bailey 1994). Potential mecha- Potter (1997) compared the diets and mouth sizes nisms leading to density-dependent predation of five estuarine species of larval fish and con- mortality as a result of predator behaviour are de- cluded that mouth width had only a small influ- scribed in Section 12.3.2. Density-dependent prey ence on prey sizes eaten and that disparate feeding responses can also influence mortality rates. For patterns among larvae were likely due to behavi- example, competition for food at high density may oural differences. Behavioural capabilities of both lead to reduced growth rates, with prey being predator and prey can also contribute considerably vulnerable to gape-limited piscivores for longer to the sizes of prey eaten. The ability of prey to periods. evade predators is related to body size, with larger prey being more efficient at avoiding predator 12.6.1 Population and strikes. Therefore, maximum prey sizes eaten community effects by piscivorous fishes can often be considerably smaller than gape sizes (Juanes and Conover 1995; Piscivorous fishes are known to have a dramatic B. Christensen 1996). influence on population- and community-level dynamics (Persson, Chapter 15, this volume). The impacts of piscivores can extend down several 12.6 POPULATION trophic levels (Zaret and Paine 1973). In freshwater REGULATION systems, the trophic cascade model describes zooplanktivore, zooplankton, and phytoplankton Shepherd and Cushing (1990) stated that abundance under high and low levels of piscivore abundance (see Carpenter et al. 1985; Persson, there is some basis for the belief that fish Chapter 15, this volume; Kaiser and Jennings, populations are regulated in some way. In Chapter 16, Volume 2). When piscivore abundance fact, the evidence is twofold: first, that they levels are high, density of zooplanktivorous fish do not explode when subjected only to low is reduced, which leads to increased zooplankton 278 Chapter 12

(herbivore) and reduced phytoplankton abundance in the USA after northern pike were introduced levels. Opposite abundance levels to those de- was emigration; this led to a rapid decline in prey scribed above are observed during periods of low fish abundance and a change in community struc- piscivore density. Hambright (1994) describes how ture (He and Kitchell 1990). the vulnerability of zooplanktivores to piscivory determines the magnitude of zooplanktivory and thus its effect on lower trophic levels. Piscivorous 12.7 METHODS OF fish can influence the life-history strategy of their STUDYING PREDATION prey. Reznick et al. (1990) showed that under sepa- IN THE FIELD rate piscivore regimes (low vs. high predation), guppies (Poecilia reticulata) evolved different life- There are several methods used to determine the history parameters such as age at maturity, repro- effects of piscivorous fish on prey populations and ductive effort, brood size and size of offspring communities. A direct approach is to make com- (Hutchings, Chapter 7, this volume). parisons between treatments where piscivores are Cannibalism is interesting because of its poten- present or absent. These treatments may be natural tial role in regulating population levels through its or artificial. Modelling has been used successfully contribution to natural mortality; however, evi- to determine the influence of predators on prey dence for density-dependent cannibalism in non- populations (Mittelbach, Chapter 11 and Persson, captive fishes has not been discovered (Smith Chapter 15, this volume). Studies that estimate and Reay 1991). Henderson and Corps (1997) predation mortality and compare this to total found that bass (Dicentrarchus labrax) year-class prey mortality make up another group of studies. strength had a 3-year periodicity, which they Estimates of consumption rate along with knowl- believed to result from cannibalism within the edge of diet composition, prey sizes and predator nursery. abundance are used to estimate predation mortali- Animal abundance levels are known to exhibit ty. Detailed descriptions of stomach content analy- periodicity as a result of predator–prey or host–par- sis and methods to estimate consumption rates are asite cycles. In a predator–prey cycle, increases in described by Jobling (Chapter 5, this volume). predator abundance are concurrent with declines Tethers (usually monofilament fishing line) are in prey population levels; subsequent periods of used to hold prey fish in place; the fate of the prey low prey abundance can lead to predator declines fish allows one to assess relative predation inten- and prey release. We are aware of only one example sity in both space and time. Danilowicz and Sale of a predator–prey cycle for a piscivorous fish. (1999) used tethering of French grunt (Haemulon Pacific cod (Gadus macrocephalus) and herring flavolineatum) to determine when, over a diel (Clupea harengus pallasi) populations in Hecate cycle, predation intensity was highest. They found Strait, British Columbia have patterns of abun- that the risk of being eaten was lowest during diur- dance that are suggestive of such a cycle (Walters nal periods and highest during dusk and nocturnal et al. 1986). The lack of evidence for a piscivore– periods. Age-0 rainbow trout (Salmo gairdneri) prey cycle may result from the flexible feeding were tethered in British Columbia lakes to deter- behaviour of piscivores, allowing them to remain mine the spatial and temporal patterns of pis- at stable population levels when one of their prey civory; it was concluded that the risk of predation populations is at low abundance. varied greatly over both space and time (Post et al. Piscivores can have indirect non-lethal effects 1998). Clearly, tethering experiments, as with on prey fishes. These include all of the behavioural other predator–prey studies, would need to con- responses made by prey when confronted with a form to the animal ethics regulations of the coun- piscivore. These behavioural responses can lead tries and institutions involved. to differences in community structure. The domi- Filming has been used to both identify preda- nant indirect effect on prey fishes in a northern bog tors and gain insight into the intensity of piscivory. Feeding Ecology of Piscivorous Fishes 279

Carr and Hixon (1995) used filming of small patch matic recent example has been the introduction of reefs to aid in identifying pelagic piscivores and for the Nile perch (Lates nilotica) to Lake Victoria, the measuring their visitation rates to experimental world’s largest tropical lake by surface area. The reefs. main effect of the predator introduction, similar to that observed in most systems where it has occurred, has been an accelerated decline of the 12.8 IMPLICATIONS FOR diverse endemic cichlid fauna of the lake and the CONSERVATION AND recent extinction of at least 200 species (Seehausen MANAGEMENT et al. 1997). However, the Nile perch introduction has also resulted in a fourfold increase in fishery Fishing alters the age and size structure of popula- yield and a doubling of fishery-related employ- tions as older and larger fish are often removed ment (Pitcher and Hart 1995; Kitchell et al. 1997). first, and eventually the fishery is supported by Piscivore stocking, particularly in freshwater the small newly recruited individuals. In many systems, has also been used as a way to improve ecosystems, large piscivorous species were the water quality in eutrophic lakes and reservoirs. initial targets of fishing. After the fishing down This process, termed ‘biomanipulation’, works of these populations, the fisheries shifted to through trophic cascading and has been quite smaller species at lower trophic levels (Pauly and successful in improving water quality of culturally Christensen, Chapter 10, Volume 2). Although eutrophic lakes and reservoirs (Drenner and most marine landings are of small planktivores, it Hambright 1999), although it can also have a is fishing and the removal of large species at high variety of unexpected negative impacts. trophic levels that affects ecosystem structure Marine reserves have been effective in restoring and functioning, leading to ‘top-down’ control and protecting many reef fishes (Polunin, Chapter (Carpenter et al. 1985; Hixon and Carr 1997; 14, Volume 2) and they appear to be a promising (Kaiser and Jennings, Chapter 16, Volume 2). For tool for marine ecosystem sustainability (National example, on Georges Bank, the decline in ground- Research Council 1999). However, reserves, if fish stocks has led to a dramatic increase in forage not planned carefully enough, can have negative fish such as herring and mackerel and a replace- impacts because of potential effects of physical ment of gadid and flounder species by small elas- factors combined with increased protection of mobranchs (Fogarty and Murawski 1998). As more piscivorous fishes. For example, in a small reserve and more stocks become overexploited, fewer top in Barbados, size and abundance of piscivores predators will dominate oceanic food webs and were greater within the reserve than on adjacent both direct and indirect community effects will be reefs, but recruitment of grunts (Haemulidae) prevalent. Predator removals can also be a deliber- was lower because of predation pressure and ate form of fisheries management, whereby de- oceanographic patterns of larval supply (Tupper creases in predation pressure are predicted to and Juanes 1999). result. However this strategy does not always proceed as predicted (V. Christensen 1996; Kaiser and Jennings, Chapter 16 and Cowx, Chapter 17, 12.9 CONCLUSIONS Volume 2). The opposite problem, with similar results, Piscivores are a diverse group of fish that show dis- occurs when predators are deliberately or inadver- tinct behaviours and specialized morphologies. tently added to ecosystems (Cowx, Chapter 17, The timing of the onset of piscivory is critical to Volume 2). The effects of piscivore introduction many species and may have led to adaptive strate- has been documented in temperate, boreal and gies that accelerate the shift to piscivory. The tropical lakes (Zaret and Paine 1973; Mills et al. predation process can be better understood by 1994; Vander Zanden et al. 1999). The most dra- examining the components of predation, and the 280 Chapter 12 predatory response to changes in prey density can Carpenter, S.R., Kitchell, J.F. and Hodgson, J.R. (1985) be estimated by generating numerical and func- Cascading trophic interactions and lake productivity. tional responses. Piscivorous fish exhibit general BioScience 35, 634–9. Carr, M.H. and Hixon, M.A. (1995) Predation effects on patterns of prey type and size selectivity driven by early post-settlement survivorship of coral-reef fishes. attack behaviours and prey-specific behavioural Marine Ecology Progress Series 124, 31–42. responses by the prey. Predatory fish can have large Chesson, J. (1978) Measuring preference in selective impacts on prey communities in all habitats where predation. Ecology 59, 211–15. they occur and are thus common targets for conser- Christensen, B. (1996) Predator foraging capabilities and vation and management, as they are often the first prey antipredator behaviours: pre- versus postcapture species to be affected by harvesting; however, they constraints on size-dependent predator–prey interac- tions. Oikos 76, 368–80. can also have both positive and detrimental effects Christensen, V. (1996) Managing fisheries involving when deliberately or inadvertently added to predator and prey species. Reviews in Fish Biology and ecosystems. Fisheries 6, 417–42. Danilowicz, B.S. and Sale, P.F. (1999) Relative intensity of predation on the French grunt, Haemulon flavolinea- tum, during diurnal, dusk, and nocturnal periods on a REFERENCES coral reef. Marine Biology 133, 337–43. Dill, L.M. (1983) Adaptive flexibility in the foraging Adams, S.M. and DeAngelis, D.L. (1987) Indirect effects behavior of fishes. Canadian Journal of Fisheries and of early bass–shad interactions on predator population Aquatic Sciences 40, 398–408. structure and food web dynamics. In: W.C. Kerfoot and Drenner, R.W. and Hambright, K.D. (1999) Biomanipula- A. Sih (eds) Predation: Direct and Indirect Impacts tion of fish assemblages as a lake restoration tech- on Aquatic Communities. Hanover, NH: University nique. Archiv für Hydrobiologie 146, 129–65. Press of New England, pp. 103–17. Eby, L.A., Rudstam, L.G. and Kitchell, J.F. (1995) Predator Bailey, K.M. (1994) Predation on juvenile flatfish and re- responses to prey population dynamics: an empirical cruitment variability. Netherlands Journal of Sea analysis based on lake trout growth rates. Canadian Research 32, 175–89. Journal of Fisheries and Aquatic Sciences 52, 1564–71. Blake, R.W. (1983) Fish Locomotion. Cambridge: Cam- Eklöv, P. (1992) Group foraging versus solitary foraging bridge University Press. efficiency in piscivorous predators: the perch, Perca Buckel, J.A., Letcher, B.H. and Conover, D.O.(1998) Ef- fluviatilis, and pike, Esox lucius, patterns. Animal fects of a delayed onset of piscivory on size of age-0 Behaviour 44, 313–26. bluefish. Transactions of the American Fisheries Soci- Fogarty, M.J. and Murawski, S.A. (1998) Large-scale dis- ety 127, 576–87. turbance and the structure of marine systems: fishery Buckel, J.A., Fogarty, M.J. and Conover, D.O. (1999a) impacts on Georges Bank. Ecological Applications 8 Mutual prey of fish and humans: a comparison of (Suppl.), S6–S22. biomass consumed by bluefish, Pomatomus saltatrix, Forney, J.L. (1971) Development of dominant year classes with that harvested by fisheries. Fishery Bulletin 97, in a yellow perch population. Transactions of the 776–85. American Fisheries Society 100, 739–49. Buckel, J.A., Conover, D.O., Steinberg, N.D. and Forrester, G.E. (1995) Strong density-dependent survival McKown, K.A. (1999b) Impact of age-0 bluefish pre- and recruitment regulate the abundance of a coral reef dation on age-0 fishes in the Hudson River estuary: fish. Oecologia 103, 275–82. evidence for density-dependent loss of juvenile striped Frankiewicz, P., Dabrowski, K. and Zalewski, M. (1996) bass. Canadian Journal of Fisheries and Aquatic Mechanism of establishing bimodality in a size distri- Sciences 56, 275–87. bution of age-0 pikeperch, Stizostedion lucioperca (L.) Buckel, J.A., Fogarty, M.J. and Conover, D.O. (1999c) For- in the Sulejow Reservoir, central Poland. Annales aging habits of bluefish, Pomatomus saltatrix, on the Zoologie Fennici 33, 321–7. U.S. east coast continental shelf. Fishery Bulletin 97, Gaughan, D.J. and Potter, I.C. (1997) Analysis of diet and 758–75. feeding strategies within an assemblage of estuarine Buijse, A.D. and Houthuijzen, R.P. (1992) Piscivory, larval fish and an objective assessment of dietary niche growth, and size-selective mortality of age 0 pikeperch overlap. Fishery Bulletin 95, 722–31. (Stizostedion lucioperca). Canadian Journal of Gerking, S.D. (1994) Feeding Ecology of Fish. New York: Fisheries and Aquatic Sciences 49, 894–902. Academic Press. Feeding Ecology of Piscivorous Fishes 281

Gerritsen, J. and Strickler, J.E. (1977) Encounter prob- piscivorous fishes? In: D.J. Stouder, K.L. Fresh and R.J. abilities and community structure in zooplankton: Feller (eds) Theory and Application in Studies of Fish a mathematical model. Journal of the Fisheries Re- Feeding Ecology. Columbia, SC: University of South search Board of Canada 34, 73–82. Carolina Press, pp. 79–100. Greene, C.H. (1986) Patterns of prey selection: implica- Juanes, F. and Conover, D.O. (1994a) Piscivory and prey tions of predator foraging tactics. American Naturalist size selection by young-of-the-year bluefish: predator 128, 824–39. preference or size-dependent capture success? Marine Hambright, K.D. (1994) Morphological constraints in Ecology Progress Series 114, 59–69. the piscivore–planktivore interaction: implications Juanes, F. and Conover, D.O. (1994b) Rapid growth, high for the trophic cascade hypothesis. Limnology and feeding rates, and early piscivory in young-of-the-year Oceanography 39, 897–912. bluefish (Pomatomus saltatrix). Canadian Journal of Hambright, K.D., Drenner, R.W., McComas, S.R. and Fisheries and Aquatic Sciences 51, 1752–61. Hairston, N.G. (1991) Gape-limited piscivores, plank- Juanes, F. and Conover, D.O. (1995) Size-structured pis- tivore size refuges, and the trophic cascade hypothesis. civory: advection and the linkage between predator Archiv für Hydrobiologie 121, 389–404. and prey recruitment in young-of-the-year bluefish. Hart, P.J.B. (1993) Teleost foraging: facts and theories. In: Marine Ecology Progress Series 128, 287–304. T.J. Pitcher (ed.) Behaviour of Teleost Fishes. London: Juanes, F., Buckel, J.A. and Conover, D.O. (1994) Acceler- Chapman & Hall, pp. 253–84. ating the onset of piscivory: intersection of predator Hart, P.J.B. and Hamrin, S.F. (1990) The role of behaviour and prey phenologies. Journal of Fish Biology 45 and morphology in the selection of prey by pike. In: (Suppl. A), 41–54. Hughes (ed.) Behavioural Mechanisms of Food Selec- Keast, A. (1985) The piscivore guild of fishes in small tion. Heidelberg: Springer-Verlag, pp. 235–54. freshwater ecosystems. Environmental Biology of Hassell, M.P. (1978) The Dynamics of Arthropod Preda- Fishes 12, 119–29. tor–Prey Systems. Princeton, NJ: Princeton University Keast, A. and Webb, D. (1966) Mouth and body form rela- Press. tive to feeding ecology in the fish fauna of a small lake, He, X. and Kitchell, J.F. (1990) Direct and indirect effects Lake Opinicon, Ontario. Journal of the Fisheries Re- of predation on a fish community: a whole lake experi- search Board of Canada 23, 1845–74. ment. Transactions of the American Fisheries Society Kitchell, J.F., Schindler, D.S., Ogutu-Ohwayo, R. and 119, 825–35. Reinthal, P.M. (1997) The Nile perch in Lake Victoria: Henderson, P. and Corps, M. (1997) The role of tempera- interactions between predation and fisheries. Ecologi- ture and cannibalism in interannual recruitment vari- cal Applications 7, 653–64. ation of bass in British waters. Journal of Fish Biology Magnússon, K.G. (1995) An overview of the multispecies 50, 280–95. VPA: theory and applications. Reviews in Fish Biology Hixon, M. and Carr, M. (1997) Synergistic predation, den- and Fisheries 5, 195–212. sity dependence, and population regulation in marine Mangel, M. (1996) Life history invariants, age at maturity fish. Science 277, 946–9. and the ferox trout. Evolutionary Ecology 10, 249–63. Hobson, E.S. (1991) Trophic relationships of fishes spe- Margulies, D. (1997) Development of the visual system cialized to feed on zooplanters above coral reefs. In: P.F. and inferred performance capabilities of larval and Sale (ed.) The Ecology of Fishes on Coral Reefs. New early juvenile scombrids. Marine Behaviour and Phys- York: Academic Press, pp. 69–95. iology 30, 75–98. Holling, C.S. (1965) The functional response of predators Mathias, J.A. and Li, S. (1982) Feeding habits of walleye to prey density and its role in mimicry and population larvae and juveniles: comparative laboratory and field regulation. Memoirs of the Entomological Society of studies. Transactions of the American Fisheries Soci- Canada 45, 3–60. ety 111, 722–35. Hoyle, J.A. and Keast, A. (1987) The effect of prey mor- Miller, T.J., Crowder, L.B., Rice, J.A. and Marschall, E.A. phology and size on handling time in a piscivore, the (1988) Larval size and recruitment mechanisms in largemouth bass (Micropterus salmoides). Canadian fishes: toward a conceptual framework. Canadian Journal of Zoology 65, 1972–7. Journal of Fisheries and Aquatic Sciences 45, 1657–70. Jenkins, G.P., Milward, N.E. and Hartwick, R.F. (1984) Mills, E.L., Leach, J.H., Carlton, J.T. and Secor, C.L. Food of larvae of Spanish mackerels, genus Scombero- (1994) Exotic species and the integrity of the Great morus (Teleostei: Scombridae), in shelf waters of the Lakes. Bioscience 44, 666–76. Great Barrier Reef. Australian Journal of Marine and Mittelbach, G.G. and Persson, L. (1998) The ontogeny of Freshwater Research 35, 477–82. piscivory and its ecological consequences. Canadian Juanes, F. (1994) What determines prey size selectivity in Journal of Fisheries and Aquatic Sciences 55, 1454–65. 282 Chapter 12

Murdoch, W.W. and Oaten, A. (1975) Predation and popu- dation on sockeye salmon smolts at Little Togiak lation stability. Advances in Ecological Research 9, River, Alaska. Fishery Bulletin 82, 401–10. 1–131. Sazima, I. (1998) Field evidence for suspension feeding in Myers, R.A. and Cadigan, N.G. (1993) Density- Pseudocaranx dentex, with comments on ram filter- dependent juvenile mortality in marine demersal fish. ing in other jacks (Carangidae). Environmental Biology Canadian Journal of Fisheries and Aquatic Sciences of Fishes 53, 225–9. 50, 1576–90. Scharf, F.S., Buckel, J.A., Juanes, F. and Conover, D.O. National Research Council (1999) Sustaining Marine (1997) Estimating piscine prey size from partial re- Fisheries. Washington, DC: National Academy Press. mains: testing for shifts in foraging mode by juvenile Olson, M.H. (1996) Ontogenetic shifts in largemouth bluefish. Environmental Biology of Fishes 49, 377–88. bass: variability and consequences for first-year Scharf, F.S., Buckel, J.A., Juanes, F. and Conover, D.O. growth. Ecology 77, 179–90. (1998) Predation by juvenile piscivorous bluefish Pearre, S. (1986) Ratio-based trophic niche breadths (Pomatomus saltatrix): the influence of prey to preda- of fish, the Sheldon spectrum, and the size–efficiency tor size ratio and prey type on predator capture success hypothesis. Marine Ecology Progress Series 27, 299– and prey profitability. Canadian Journal of Fisheries 314. and Aquatic Sciences 55, 1695–703. Persson, L., Andersson, J., Wahlstrom, E. and Eklov, P. Scharf, F.S., Juanes, F. and Rountree, R.A. (2000) Predator (1996) Size-specific interactions in lake systems: size–prey size relationships of marine fish predators: predator gape limitation and prey growth rate and mor- interspecific variation and the effects of ontogeny and tality. Ecology 77, 900–11. body size on trophic niche breadth. Marine Ecology Peterman, R.M. and Gatto, M. (1978) Estimation of Progress Series 208, 229–48. functional responses of predators on juvenile salmon. Seehausen, O., Witte, F., Katunzi, E.F., Smits, J. and Journal of the Fisheries Research Board of Canada 35, Bouton, N. (1997) Patterns of remnant cichlid fauna 797–808. in southern Lake Victoria. Conservation Biology 11, Petersen, J.H. and DeAngelis, D.L. (1992) Functional 890–904. response and capture timing in an individual-based Shepherd, J.G. and Cushing, D.H. (1990) Regulation in model: predation by northern squawfish (Pty- fish populations: myth or mirage? Philosophical chocheilus oregonensis) on juvenile salmonids in the Transactions of the Royal Society of London Series B Columbia River. Canadian Journal of Fisheries and 330, 151–64. Aquatic Sciences 49, 2551–65. Sih, A. and Christensen, B. (2001) Optimal diet theory: Pitcher, T.J. and Hart, P.J.B. (1995) The Impacts of Species when does it work and when and why does it fail? Changes in African Lakes. London: Chapman & Animal Behaviour 61, 379–90. Hall. Skulason, S., Snorrason, S.S., Ota, D. and Noakes, D.L.G. Pitcher, T.J. and Parrish, J.K. (1993) Functions of shoaling (1993) Genetically based differences in foraging behav- behaviour in teleosts. In: T.J. Pitcher (ed.) The Behav- iour among sympatric morphs of arctic charr (Pisces: iour of Teleost Fishes. London: Chapman & Hall, pp. Salmonidae). Animal Behaviour 45, 1179–92. 363–439. Smith, C. and Reay, P. (1991) Cannibalism in teleost fish. Post, J.R., Parkinson, E.A. and Johnston, N.T. (1998) Spa- Reviews in Fish Biology and Fisheries 1, 41–64. tial and temporal variation in risk to piscivory of age-0 Sogard, S.M. (1997) Size-selective mortality in the juve- rainbow trout: patterns and population level conse- nile stage of teleost fishes: a review. Bulletin of Marine quences. Transactions of the American Fisheries Science 60, 1129–57. Society 127, 932–42. Stephens, D.W. and Krebs, J.R. (1986) Foraging Theory. Reimchen, T.E. (1991) Evolutionary attributes of head- Princeton, NJ: Princeton University Press. first prey manipulation and swallowing in piscivores. Tanaka, M., Kaji, T., Nakamura, Y. and Takahashi, Y. Canadian Journal of Zoology 69, 2912–16. (1996) Developmental strategies of scombrid larvae: Reznick, D.A., Bryga, H. and Endler, J.A. (1990) Experi- high growth potential related to food habits and preco- mentally induced life-history evolution in a natural cious digestive system development. In: Y. Watanabe, population. Nature 346, 357–9. Y. Yamashita and Y. Oozeki (eds) Survival Strategies in Rice, J.A., Crowder, L.B. and Rose, K.A. (1993) Interac- Early Life Stages of Marine Resources. Rotterdam: tions between size-structured predator and prey popu- A.A. Balkema, pp. 125–39. lations: experimental test and model comparison. Tupper, M. and Juanes, F. (1999) Effects of a marine re- Transactions of the American Fisheries Society 122, serve on recruitment of grunts (Pisces: Haemulidae) 481–91. at Barbados, West Indies. Environmental Biology of Ruggerone, G.T. and Rogers, D.E. (1983) Arctic charr pre- Fishes 55, 53–63. Feeding Ecology of Piscivorous Fishes 283

Utne-Palm, A.C. (2000) Prey visibility, activity, size and Werner, E.E. (1977) Species packing and niche comple- catchability’s (evasiveness) influence on Gobiusculus mentarity in three sunfishes. American Naturalist flavescens prey choice. Sarsia 85, 157–65. 111, 553–78. Vander Zanden, M.J., Casselman, J.M. and Rasmussen, Werner, E.E. and Gilliam, J.F. (1984) The ontogenetic J.B. (1999) Stable isotope evidence for the food web con- niche and species interactions in size-structured popu- sequences of species invasions in lakes. Nature 401, lations. Annual Review of Ecology and Systematics 464–7. 15, 393–425. Wahl, D.H. and Stein, R.A. (1988) Selective predation by Wilson, D.S. (1975) The adequacy of body size as a niche three esocids: the role of prey behavior and morphol- difference. American Naturalist 109, 769–84. ogy. Transactions of the American Fisheries Society Winemiller, K.O. (1989) Ontogenetic diet shifts and re- 117, 142–51. source partitioning among piscivorous fishes in the Wainwright, P.C. and Richard, B.A. (1995) Predicting pat- Venezuelan Llanos. Environmental Biology of Fishes terns of prey use from morphology of fishes. Environ- 26, 177–99. mental Biology of Fishes 44, 97–113. Wootton, R.J. (1990) Ecology of Teleost Fishes. New York: Walters, C.J., Stocker, M., Tyler, A.V. and Westrheim, S.J. Chapman & Hall. (1986) Interaction between Pacific cod (Gadus macro- Wright, R.A., Crowder, L.B. and Martin, T.H. (1993) cephalus) and herring (Clupea harengus pallasi) in the The effects of predation on the survival and size- Hecate Strait, British Columbia. Canadian Journal of distribution of estuarine fishes: an experimental Fisheries and Aquatic Sciences 43, 830–7. approach. Environmental Biology of Fishes 36, 291– Webb, P.W. (1976) The effect of size on the fast-start 300. performance of rainbow trout, Salmo gairdneri, and Zaret, T.M. and Paine, R.T. (1973) Species introduction in a consideration of piscivorous predator–prey interac- a tropical lake. Science 182, 449–55. tions. Journal of Experimental Biology 65, 157–77. 13 Fish as Prey

J. KRAUSE, E.M.A. HENSOR AND G.D. RUXTON

13.1 INTRODUCTION behaviour by prey might be to find out whe- ther or not predators are hungry (Dugatkin and Most fish experience some degree of predation Godin 1992). Licht (1989) has shown that guppies, risk throughout their lives. There are numerous Poecilia reticulata, are indeed capable of making strategies that reduce this risk. Predation has this distinction. such an immediate effect on fitness that it is a On encountering a predator, a fish can either powerful selective force, and different species freeze or flee (Fig. 13.1), and split-second decisions from different environments have reached a vari- made by a prey fish can greatly affect its chances of ety of solutions to reduce predation risk (Helfman survival. These decisions must be based on a num- et al. 1997). Although evasion of predators is ber of factors, such as distance to the nearest cover, important, fish cannot hide indefinitely. They the physical condition of the fish, the potential must also do some or all of feeding, courting, efficacy of crypsis, which depends on light levels mating, spawning and defending territories and substratum type, or degree of protection from against competitors, all of which are activities spines and other morphological predator deter- important to an individual’s fitness. Therefore fish rents. Fish with well-developed defences may be must accept some degree of predation risk, but can more likely to opt for immobility, relying on minimize the danger by efficiently detecting and those defences to protect them if they are detected reducing their vulnerability to predators (Godin and attacked. Fish with fewer morphological 1997; Smith 1997). antipredator adaptations might be more likely to One of the most important tasks of prey is to flee from danger. In reality, however, fish may need correctly identify what does and does not pose a to combine a number of strategies to achieve the potential predation threat. Not all large species optimal escape mechanism. are necessarily dangerous and there should be a In this chapter we highlight the various methods high selection pressure on being able to recognize that prey fish employ to reduce their vulnerability predators. Karplus and Algom (1981) showed that to predators. We focus on those areas of prey de- prey are indeed capable of distinguishing between fences that have not received much attention the facial features of predatory and non-predatory lately in terms of books and reviews but which reef fish species. Once a predator has been identi- have been the subject of substantial recent re- fied, it is important for fish to obtain information search. Therefore we touch lightly on morpholo- on whether an attack is likely to take place or not. gical defences, such as crypsis, body armour and This will depend largely on whether the predator is alarm substance, and refer to recent detailed treat- hungry or satiated. It has been suggested that one ments of these topics, whereas recent investiga- of the important functions of predator inspection tions of the size and composition of free-ranging Fish as Prey 285

Encounter 13.2.1 Avoiding identification as prey Immobility Mobility Crypsis Crypsis allows prey to avoid detection by the Body armour predator if it blends in with the background. This is Electric shocks a strategy found in many transparent fish larvae, Chemical stimuli counter-shaded pelagic fish and flatfish (Helfman et al. 1997; Smith 1997). Crypsis also works if it mimics objects that are detected by the predator Social refuge (shoaling) Spatial refuge (hiding) but not perceived as prey (Endler 1986). It usually conspecifics temporal refuge works best if the prey is immobile because most heterospecifics structural refuge familiar individuals predator species are very sensitive to movements and often select their prey on this basis (Krause and Fig. 13.1 Summary of the various strategies fish employ Godin 1995). In some cases such as the leaf-fish, to reduce the risk of predation following an encounter Monociorrhus polyacanthus, slow movements with a predator. resembling that of a drifting leaf are compatible with crypsis (Breder 1946; Randall and Randall 1960; Randall & Emery 1971). However, even if slow movement remains possible, behaviours shoals and experimental work on refuge use are such as foraging, mating and territorial defence given more space. remain largely incompatible with crypsis . Batesian mimicry is where a species mimics another species that is distasteful or dangerous to 13.2 IMMOBILITY the predator and is another strategy used by fishes to escape being identified as palatable. This is Immobility only works as an antipredator behav- thought to occur in a number of fish species. iour if a fish is well camouflaged, has some mor- The venomous blenny, Meiacanthus atrodorsalis, phological adaptation to make it unpalatable or is apparently mimicked by non-venomous can be confident the predator’s senses are too poor heterospecifics (Losey 1972), and a cardinalfish to detect it when still. Aspects of prey fish mor- (Apogonidae) resembles a venomous scorpionfish phology that play a part in predator avoidance and (Scorpaenidae) (Seigel and Adamson 1983). An- deterrence include crypsis, body armour, changes other attack deterrent is the production of eye- in body shape, electric shocks and the production spots that mimic the large eyes of predatory of distasteful substances or toxins. Spines, chro- species (Winemiller 1990). One of the most strik- matophores, electrophores and extra body mass ing examples of this kind of mimicry is the moray are all costly to produce at the outset but can later eel, Gymnothorax meleagris, which is mimicked benefit the fish in energetic terms, in addition by Calloplesiops altivelis that displays a con- to the antipredator benefits conferred. If the fish is spicuous eyespot when faced with a potential sufficiently well protected to afford immobility predator (McCosker 1977). The occurrence of in the face of predation risk, it saves the energy Batesian mimicry in the fish world is well sup- that more vulnerable fish expend in fleeing from ported but so far little direct evidence exists for the predator and seeking shelter. Remaining in Müllerian mimicry, in which toxic or distasteful open water has the added advantage of allowing species have coevolved the same bright warning the fish better access to information about the colours to advertise their unpalatability to mutual whereabouts of the predator and enabling judge- benefit. The latter is relatively common in insects ments about when it is safe to resume normal (Ritland 1995). This could be an interesting area for activity. further research. It is especially worth bearing in 286 Chapter 13 mind that some forms of mimicry in fish might not 1992). In response to the presence of pike, Esox be readily apparent to humans, because it is based lucius, crucian carp are capable of substantially in- on a sense other than vision. creasing their body depth within a period of several weeks. Deeper bodies increase handling time for 13.2.2 Avoiding subjugation gape-limited pike and increase the chances of es- after capture cape for prey (Nilsson et al. 1995). The increase in body depth has also been shown to be reversible and Once a fish has been captured by a predator there is thus seems to represent an adaptation to short-term still a chance of escape. A number of morphologi- invasions of crucian carp habitat by pike. cal and behavioural adaptations have evolved to Many fish species are known to release an alarm effect this. Several so-called mechanical methods substance from specialized club cells when their (sensu Endler 1991) can potentially prevent the skin is torn during handling by a predator (Smith predator from swallowing its prey. In many popu- 1997). On release, this substance can trigger a lations of three-spine sticklebacks, Gasterosteus strong alarm response in conspecifics and hetero- aculeatus, fish have evolved long dorsal and ven- specifics. Observation of this phenomenon has tral spines and bony plates (Reimchen 1994) that led to speculation that the primary function of make it difficult for many predatory species to this substance may be to provide a warning signal. swallow them (Hoogland et al. 1957). Similar adap- However, there would only be a benefit to the tations can be found in porcupinefish and puffer- sender if individuals within shoals were closely fish, which inflate their bodies and erect spines related to each other. The first observation on (Brainerd 1994; Wainwright et al. 1995). Reimchen alarm substance was made by von Frisch (1941) on (1994) pointed out that prey species often have minnows, Phoxinus phoxinus, for which no evi- multiple predators and therefore have to adapt to a dence of close relatedness between shoal members multitude of different selection pressures, result- has been found (Naish et al. 1993). It has been ing in morphological diversity. Some populations demonstrated in a number of species, such as of three-spine sticklebacks show a high degree of three-spine stickleback, that individuals are capa- diversity in adult body size. Larger body size re- ble of discriminating between sibs and non-sibs. duces predation risk from gape-limited predators, However, evidence from the field for preferential whereas small adult size allows earlier reproduc- association with kin is scarce (Ferguson and tion at the cost of higher predation (for further Noakes 1981; Poyard et al. 1999; see Krause et al. discussion of the timing of reproduction and 2000 for a review). Most of the studies on alarm mortality, see Hutchings, Chapter 7, this volume). substance have been conducted in the laboratory The coexistence of such different life-history and evidence of its effectiveness in the field is strategies depends on whether there is a fitness weak. Recent experiments showed that minnows equilibrium for a mixed evolutionarily stable failed to respond to alarm substance under field strategy (Parker 1984). Furthermore, there is good conditions, or showed only a very weak response, evidence that traits such as spine length and num- suggesting that the responses of fish to alarm sub- ber of lateral plates are under strong selection pres- stance may be highly context-dependent (Magur- sure by predators. Moodie (1972), for instance, ran et al. 1996; Irving and Magurran 1997). A study showed that the spine length of sticklebacks that by Mathis et al. (1995) demonstrated that predators were consumed by cutthroat trout, Oncorhynchus are actually attracted by alarm substance and has clarki, were proportionately shorter than those in led to the idea that the function of the substance is the general population. to attract additional predators, which will inter- A particularly interesting form of inducible fere and thus increase the prey’s chances of escape. morphological defence has been reported in crucian The processes leading to the evolution of alarm carp, Carassius carassius (Brönmark and Miner substance in fish could thus be analogous to those Fish as Prey 287 underlying distress calls in birds and mammals. groups could be linked to the fact that larger groups Further research is needed. are more conspicuous. In fact, predator preference is related specifically to the activity level of groups and not merely their size. Krause and Godin 13.3 MOBILITY (1995) manipulated shoal activity in guppies, P. reticulata, by using water of different tempera- 13.3.1 Grouping tures to change the activity level of fish and showed that blue acara cichlids, Aequidens pul- Group size cher, the predator, preferred the most active groups Shoaling has been shown to provide effective irrespective of group size. Turner and Pitcher protection from a number of predator species (1986) illustrated the important role of prey con- (Neill and Cullen 1974; Major 1978; Landeau and spicuousness by showing that individuals are safer Terborgh 1986; Krause and Godin 1995). Neill and in groups than alone, provided that the probability Cullen (1974) exposed prey to four species of pre- of detection and subsequent attack by predators dators: two fish predators (perch, Perca fluviatilis, does not increase over-proportionately with in- and pike) and two cephalopod species (squid, creasing group size. Predators can take at most one Loligo vulgaris, and cuttlefish, Sepia officinalis). individual in an attack, a phenomenon that has In the experiments with fish predators, juvenile been termed the ‘encounter-dilution effect’ (see bleak (Alburnus alburnus), dace (Leuciscus leucis- Pitcher and Parrish 1993 for details). cus) and minnow, and cyprinodontids (Poecilia Interestingly, grouping does not always protect reticulata) were used as prey. For the cephalopod prey from predators. In a study of shoals of creek predators, atherinids (Atherina spp.) and mullets chub, Semotilus atromaculatus, attacked by rock (Mugil spp.) were used. The experiments demon- bass, Ambloplites rupestris, Krause et al. (1998a) strated that prey per-capita predation risk de- found no effect of group size on per-capita preda- creased dramatically with increasing group size in tion risk if predators managed to surprise prey each case. There are a number of group size-related groups. This was attributed to the fact that effects besides simple dilution that are thought to surprise attacks may at least partially circumvent be responsible for this decrease in risk. The pres- antipredator strategies by prey. It is possible, for ence of multiple moving targets can confuse pre- instance, that the confusion effect is weakened if dators, making it difficult to target one individual predators are not detected prior to an attack. Per- (Neill and Cullen 1974; Milinski 1977; Krakauer haps because their large gape allows them to 1995). This reduces the chances of an attack being consume entire shoals at once, several species of initiated as well as actual capture success if an cetacean predators are also known to be unaffected attack is made. Another important factor that by prey group size and in some cases they even take makes life in groups safer is the presence of mul- advantage of grouped prey (Nottestad 1999). The tiple watchful individuals, making it harder for sonar that many large aquatic predators use to find predators to approach undetected and stage sur- their food may require shoals to be above a certain prise attacks (Magurran et al. 1985). size in order to be detected. However, shoaling is a Despite the fact that predator hunting success common strategy among prey fish targeted by such decreases, in some situations larger groups are predators. This has to be seen in the context of an preferentially attacked compared with smaller environment that has multiple predators, which ones (Krause and Godin 1995). However, the pref- makes it impossible for prey to simultaneously erence for larger groups is not strong enough to adopt optimal strategies against each type of pre- override the above antipredator benefits of dator. The assumption here is that cetaceans are grouping so that per-capita risk still declines with not the major predators of such species and ac- group size. The fact that predators prefer larger count for only a small proportion of the overall 288 Chapter 13 mortality rate due to predation, whilst other more benefit from grouping with a highly abundant important predators select for large group sizes. species B by being better protected against pre- dators in larger groups. However, for this benefit to be fully realized, individuals of species A would Group composition have to resemble individuals of species B. Other- There is general consensus that the presence of wise they would appear as phenotypically odd and predators in the environment is one of the main the advantage of joining a larger group might be causes for the evolution of grouping in prey and an outweighed by being odd in the group. In popula- important determinant of prey group size (Pulliam tions where a certain overlap of phenotypic charac- and Caraco 1984; Godin 1986, 1997; Magurran teristics between different species already exists, a 1990; Pitcher and Parrish 1993; Helfman et al. preference of predators for odd prey could thus 1997). However, predators also have more subtle become the driving force behind the evolution of influences on the grouping strategy of prey. Some mimicry rings that could potentially involve a large predator species suffer from the confusion effect of number of different species (Ehrlich and Ehrlich grouping prey and one way of reducing this prob- 1973). The above-mentioned group-size effect lem is to focus on prey that differ phenotypically would also benefit the mimicked species B, thus from the rest of the group (Landeau and Terborgh guaranteeing the stability of such mimicry rings. A 1986; Theodorakis 1989). The phenomenon first example of this form of mimicry has been whereby predators single out such prey individuals described by Dafni and Diamant (1984) who found a is called the oddity effect. It is typical of predators system in which a solitary fish species mimics a that attack mobile grouping prey, whereas preda- shoaling species to gain protection from grouping. tors that feed on stationary prey usually develop a However, there may be circumstances where the prey search image based on the phenotype of the mimicked species B does not benefit from the most common prey type (Beukema 1968). Thus mimicry, perhaps because predators are less able to there are two frequency-dependent types of preda- learn an aposematic signal. The importance of tion: one that is positively frequency-dependent, mimicry to fish is a promising area of research. where the predator prefers the common prey The common phenotype is only preferred by type, and another which is negatively frequency- predators that are unconstrained by the confusion dependent, where the predator prefers the odd prey effect. This should lead to the evolution of poly- type. These two types of predation select for differ- morphic prey groups because any prey phenotype ent grouping tendencies in prey. If predators prefer that dominates the group numerically will be odd prey, then we should expect phenotype match- selected against. Also, predators encountering a ing to be selected for. Indeed the latter has been diversity of different escape strategies will find it demonstrated in a number of studies: in laboratory more difficult to learn how to most effectively experiments individual banded killifish, Fundulus combat any one of them. diaphanus, preferred to shoal with conspecifics over heterospecifics, and three-spine stickleback Group geometry and two-spotted goby, Gobiusculus flavescens, preferred to shoal with size-matched over un- So far the assumption has been that an individual matched individuals; these fish were also found to outside a group is subject to a high predation risk be assorted by species and body length in the field and that all individuals within a group have the (Ranta et al. 1992; Krause 1994; Krause and Godin same lower predation risk. The latter assumption, 1994; Krause et al. 1996; Peuhkuri et al. 1997; however, is rather unlikely. Hamilton (1971), in Svensson et al. 2000). his ground-breaking study on the geometry of The preference of predators for phenotypically selfish herds, predicted that individuals on the odd prey could potentially lead to the evolution of edge of a group have a higher risk of predation than mimicry. A less abundant species A could gain a centrally positioned ones. This prediction was Fish as Prey 289 based on the assumption that the predator always in the front part of shoals and an elongated overall attacks the nearest prey within range. Hamilton’s shape were met (Bumann et al. 1997). theory of marginal predation is well supported by evidence from stationary animal groups (Krause 1994) but there is controversy over whether it can 13.3.2 Refuge use be applied to mobile fish shoals (Parrish 1989; Spatial and temporal refugia Krause 1993; Bumann et al. 1997; Krause et al. 1998a). Most studies to date have only provided The most familiar form of spatial refuge is struc- indirect evidence for or against Hamilton’s theory tural, for example a crevice under a rock or a patch by measuring the number of attacks directed at of weed. In pelagic zones, however, there may not certain shoal positions (Parrish 1989) or by looking be fixed visible refuges. The water column may be at positioning behaviour of fish (McKaye and divided into areas of different depths or tempera- Oliver 1980; Krause 1993). Some of the studies tures, which still offer shelter to prey. For example, that recorded actual mortalities employed vague sufficient depth offers protection from predators definitions of shoal positions or did not correct that use vision to locate prey. Many fish-eating for the fact that numbers of individuals in the birds are restricted to the top few metres of the centre and periphery of a group might have been water. Streams can contain shallow fast-moving different (Parrish et al. 1989). In a more recent riffles and deeper slow-moving pools, each having study, Bumann et al. (1997) extended Hamilton’s their own distinct characteristics with respect to model to mobile groups and stationary as well as predation risk (Bremset and Berg 1999). mobile predators, making predictions about the Bremset and Berg (1999) observed juvenile distribution of predation risks within fish shoals. trout, Salmo trutta, and Atlantic salmon, S. salar, The results from the model indicate that in addi- of different size classes and found that smaller tion to the positive risk gradient from the centre to juveniles stationed themselves closer to the river the periphery, as predicted by Hamilton’s model bank, nearer the bottom than larger juveniles. for stationary groups, there is another positive risk Food availability was higher in the open water; gradient from the relatively more safe rear to the however the predation risk also increased with dis- relatively less safe front part of groups. This predic- tance from the river bank. Thus young-of-the-year tion was tested using a shoal of 13 creek chub as salmonids were excluded from the upper part of prey and rock bass as a predator. Chub in the front the water column, in part by competition from half of the shoal were attacked and captured larger more dominant conspecifics and partly by a significantly more often than ones in the rear, and need to reduce the risk of predation and/or canni- the lead fish was the one with the highest pre- balism by older conspecifics. In this case, the dation risk in the shoal (Krause et al. 1998a). The smaller fish restricted their movements to the existence of risk gradients within fish shoals is river bank and bottom only of the river as a whole, likely to influence individual positioning behav- and therefore could be considered to have been iour and thus may have consequences for overall using these areas as a spatial refuge. group shape and structure. If the risk is higher for Prey can reduce their predation risk by being front positions, we may expect that the inter- active when predators are absent or the likelihood individual distances between fish will be lower of successful attack is reduced. The cycles of emer- there as a means of risk reduction (sensu Hamilton gence seen in the periodical cicada (Magicicada 1971). We might also expect that fish in middle and spp.) are thought to play a part in predator avoid- rear positions ‘hide’ behind frontal fish and thus ance (Itô 1998), as are the diurnal patterns of inver- give the shoal an elongated structure (Bumann et tebrate drift found in some freshwater systems al. 1997). There is an indication from detailed (McIntosh and Peckarsky 1996; Huhta et al. 1999). three-dimensional measurements of roach shoals Behavioural periodicity related to predation risk that both the predictions for higher fish densities has also been demonstrated in some fish species. 290 Chapter 13

Table 13.1 Costs and benefits of refuge use.

Costs Benefits

Lost opportunity Safety from Feeding (Dill and Fraser 1997) Immediate predation (Milinski 1993; Sogard and Olla 1993; Smith Courting (Perrin 1995) 1997) Mating (Perrin 1995) Predation when vulnerable (e.g. when spawning/brooding/newly hatched) (Sogard 1997; Rodewald and Foster 1998) Uncertainty Food availability (Sih 1992, 1997; Dill and Fraser 1997) Metabolic effects Predator presence (Sih 1992; Eklöv and Persson 1996) Reduced energy expenditure when food supply poor (Dill and Fraser 1997) Social effects within refuge Increased competition for space, position, information and food Protection from (Pallini et al. 1998) Damaging environmental conditions, e.g. currents, extreme temperatures (Brown 1999) Social effects outside refuge Reduced defence of territory (Sargent 1985), food (Humphries et al. Reduced stress 1999) and mates (Sargent 1985)

For example, Naud and Magnan (1988) studied 11, this volume. We therefore provide only a cur- evening offshore migrations by northern redbelly sory review of costs of refuge use here. The extent dace, Phoxinus eos, and concluded that during the to which an individual benefits from the reduced day the fish were forced to form shoals near physi- predation risk will vary according to the losses cal structures to reduce their risk of predation by it incurs while in hiding. Juvenile fish, often brook charr, Salvelinus fontinalis, thus forgoing comparatively more at risk of predation than the opportunity to feed in the more productive larger conspecifics (Sogard 1997; Krause et al. limnetic habitat. Some marine fish larvae have 1998b), may trade off reduced feeding opportunity been shown to start their drift up into the water for increased protection from predators by remain- column when the prevailing currents ensure food ing in safer but less productive areas (Bremset is plentiful in the surface waters and predation by and Berg 1999). Conversely, individuals in poorer macrozooplankton is at a minimum (Frank and condition or with higher nutritional require- Leggett 1985). A seasonal switch from diurnal to ments, perhaps as a result of parasitism, may be nocturnal foraging at low temperatures has been prepared to take more risks in order to maximize observed in juvenile Atlantic salmon (Fraser et al. their food intake (Giles 1983; Milinski 1990, 1993). 1993; Fraser and Metcalfe 1997; Metcalfe et al. Growing large enough to become invulnerable to 1999) and minnows (Greenwood and Metcalfe some types of predator can also be seen as entering 1998). Metcalfe et al. (1999) concluded that by in- a partial refuge. The various costs and benefits of creasing their nocturnal activity relative to their refuge use, any of which a fish may have to balance diurnal activity, the salmon experienced reduced during hiding, are shown in Table 13.1. predation pressure from visual hunters, at a time when their nutritional requirements were low The costs of refuge use enough to allow reduced diurnal foraging. Lost opportunity The opportunities to feed, court or mate while in hiding may be greatly reduced Refuge trade-offs (Sih 1992). In the case of spatial refugia, food Trade-offs, particularly between foraging and pre- availability and the chances of interaction with dation risk, are reviewed by Mittelbach in Chapter conspecifics are potentially lower (Mittelbach, Fish as Prey 291

Chapter 11, this volume). Fish that seek temporal and then refeeding them again over 2 days, Krause refuge, by avoiding activity during risky periods, et al. (1999) demonstrated an inverse relationship may be missing the best times for feeding or mating. between percentage weight lost and hiding dura- For example, Fraser and Metcalfe (1997) concluded tion. Hungrier fish seemed to accept a higher that the feeding efficiency of juvenile Atlantic predation risk in return for increased foraging salmon was significantly impaired by low light opportunity. levels during their switch to nocturnal feeding in By assessing the conditions outside the refuge, winter. This was not compensated for by winter fish can weigh up the advantages of remaining drift abundance being slightly higher by night than inside. Information about food availability or pre- by day. They concluded that many individuals ac- dator presence gleaned from within the refuge will cepted the minimum intake required for survival, form the basis of a decision to leave. However, fish in return for a significantly lower risk of predation are not omniscient (Sih 1992). Underestimating by predominantly diurnal avian predators. the availability, or overestimating the quality, of The extent of the opportunity lost depends on food may lead to increased searching or handling the nature of the resource. Food availability in time, thereby increasing the risk of detection by aquatic environments is seldom constant: a fish a predator (Sih 1992). Given the choice between relying on a patchy food resource may be forced to death, the immediate cost of underestimating the take advantage of a plentiful supply whenever it predation risk, and lost feeding opportunity, which arises, as the arrival of the next meal cannot be comes from overestimating the predation risk, a predicted. Dill and Fraser (1997) demonstrated that few animals will sometimes stay in hiding much polychaete worms, Serpula vermicularis, emerged longer than actually necessary to avoid the sooner after a fright stimulus when the concentra- predator (Sih 1997). tion of food present at the time of hiding was greater. They concluded that the worms were Social effects within the refuge Fish that seek tracking short-term fluctuations in local food shelter together experience different social costs. availability and were modifying their hiding dura- Whilst hiding in a spatial refuge they may have to tion accordingly. Since food availability outside compete for resources inside, which are in any case the refuge can vary greatly, the costs incurred by likely to be limited. Information about the condi- remaining in hiding also change over time. If there tions outside the refuge affects survival and can be is no food available, an individual can save energy considered a resource, in which case the spatial by remaining in the refuge. position of a fish within a refuge may affect its A fish that is better able to track fluctuations in fitness. A fish closer to the boundary with the its food resource will operate closer to the optima wider habitat will be in a better position to assess for refuging and foraging and will thus improve its the conditions outside the refuge than one further fitness. However, the accuracy with which fish back, although this position may leave it margin- can assess the external conditions may be im- ally more at risk of being spotted by a predator. paired, because remaining in the refuge restricts Observation of the fate of individuals entering or the amount of information available. This leads to leaving the refuge can enable estimation of the the problem of uncertainty. local predation risk. Allied to this, there may be antipredatory benefits in leaving the refuge in a Uncertainty The condition of the fish, such as its group. However, fish may experience increased nutritional or reproductive status, will have a bear- competition if many individuals emerge together ing on the amount of time it can spend in hiding at a particular time of day. before it must feed or commence other activities. For example, a hungry fish may be prepared to Social effects outside the refuge By seeking leave hiding sooner (Krause et al. 1998b, 1999). By shelter, a fish may be unable to defend its territory, depriving sticklebacks of food over a 2-day period, food sources or mates from competitors. The costs 292 Chapter 13 of this depend largely on whether its competitors Predation y = 1.15 + 0.04x are equally at risk from predation and how many 3.5 Control y = 0.32 + 0.05x competitors are in the vicinity. Competitors may Food deprivation y = –0.64 + 0.07x P employ similar refuge strategies, and could also be C 3.0 in hiding, or they may have their own territory to F defend. When a fish is forced away from guarding a 2.5 food resource, its potential loss in favour of its competitors depends on their feeding rate and how far they have to venture from their own territory. 2.0 In some situations, a vacated territory can be 1.5 quickly won back on return from refuge, although Log time (s) this will not be universally true. Some systems show a strong ‘owner’s advantage’, where it is 1.0 easier to hold on to a territory than it is to usurp another. Such effects can also affect third parties. 0.5 If a territory holder finds that its neighbour has taken refuge and been replaced by another, then it 0.0 is likely to experience a higher rate of aggression 18 28 38 48 from the new neighbour than previously from the Body length (mm) familiar ‘dear enemy’. There are also reproductive costs to a fish in Fig. 13.2 Experiment by Krause et al. (1998b). Individual sticklebacks (n = 30) of different body size hiding. The male three-spine stickleback must were released and took shelter under a polystyrene spend considerable time and effort tending his nest float. The behaviour was observed under three different if his eggs are to hatch successfully. He must defend treatments: control (C), simulated predator attack (P) the nest from raids by conspecifics and must fan the and food deprivation of 2 days (F). The time of emergence eggs to prevent them from dying through lack of showed a significant increase with body size in all three 2 oxygen (Wootton 1976). He cannot spend too much treatments (control, r = 0.84, P < 0.001; predation, 2 2 time away from the nest engaged in antipredator r = 0.71, P < 0.001; deprivation, r = 0.83, P < 0.001). behaviour, resting or feeding (Blouw 1996).

Juvenile fish are generally, although not always, The benefits of refuge use at higher risk of predation (Sogard 1997). However, Safety from predation One would expect indivi- it does not always follow that a higher risk of preda- duals that experience higher predation risk to seek tion will lead to longer hiding times. Small juve- refuge more often or for longer periods of time. For nile fish have a high relative metabolic rate and are instance, Rodewald and Foster (1998) found that therefore more at risk from starvation than larger gravid three-spined sticklebacks remained in fish. In fact, Krause et al. (1998b) observed that closer proximity to refuge than non-gravid females large fish emerged later from a refuge and returned in lakes harbouring predatory fish. This difference to the refuge sooner than small fish. They con- in behaviour was not observed in lakes where cluded that the risk of starvation decreased more predatory fish were absent. Carrying eggs makes sharply than the risk of predation as the fish in- the female stickleback’s body distended, and she creased in length (Fig. 13.2). Therefore, larger fish suffers a reduction in agility and stamina as a result could better afford the lost feeding costs of (Rodewald and Foster 1998). This reduction in remaining in hiding than smaller conspecifics, escape ability was concluded to be the most likely which were forced to leave refuge in search of food. explanation for the gravid females remaining The large number of factors affecting refuge use closer to refuge. makes it hard to predict how long a particular Fish as Prey 293 individual is likely to remain in hiding. Krause from a reduction in stress levels. They do not have et al. (1999) showed that emergence times in G. to be as vigilant for predators as they are outside, aculeatus, although greatly variable between each unless assessing the predation risk before leaving fish, were highly consistent for individual fish the refuge. between different days of testing. This was likely to have been due to individual differences in weight changes over the 2-day period of food depri- 13.4 CONCLUSIONS vation. Further examples of how refuges are used and how use changes with age, giving an ontoge- Fish have evolved a wide range of behavioural and netic niche shift, are given by Mittelbach (Chapter morphological adaptations to reduce their suscep- 11, this volume). tibility to predators. However, the intrinsic prob- lems faced by prey fish are similar in all cases. Protection from adverse conditions Spatial and Generally, an individual must assess the relative temporal refugia also provide protection from profitability of an antipredator strategy in terms of strong currents, extreme temperatures and other the associated costs and benefits with regards to potentially damaging environmental conditions. factors such as predation risk, access to informa- The temporal refuge sought by juvenile Atlantic tion, lost opportunities and energy expenditure. In salmon at low temperatures, when they restrict this chapter we have described how individuals their diurnal foraging, is one example. The usual can remain immobile, relying on not being identi- method of attack avoidance that salmon employ is fied or on being able to escape after capture or by by a burst acceleration towards cover. However, forming groups or taking refuge, and so reduce because both reaction time and acceleration rate their costs of encounters with predators. Rela- are temperature dependent, at low temperatures tively little information currently exists on how the ability of the fish to escape is severely im- individuals select one particular strategy from a paired, leaving them at much higher risk from whole set of available ones, how repeatable such endothermic predators (Fraser et al. 1993; Fraser choices are and on the thresholds at which dyna- and Metcalfe 1997; Metcalfe et al. 1999). Brown mic strategy changes might take place. For in- (1999) noted that cutthroat trout in Dutch Creek, stance, is there a group size at which grouping Alberta formed aggregations during the winter becomes the preferred antipredator strategy over months and that the majority then remained seden- refuge use? Grouping might not be as safe as tary until temperatures increased again. However, refuging but it allows for continued foraging. The some fish were forced to move to warmer stretches trade-off between predation risk and foraging of the river when anchor ice began to form in benefits should also depend on the nutritional the water column. Anchor ice develops from the state of the individual, making grouping particu- bottom of a pool upwards and can choke it to the larly attractive to hungry fish. extent that the water flow is directed down a few Survival depends on swift and accurate high-velocity channels, in which the fish find it decision-making by prey during encounters with hard to maintain position (Brown 1999). They must predators and little is presently known on poten- then seek refuge in warmer water. tial constraints regarding the speed of information processing and strategy choice in this context. Reduced stress Stress in fish can be caused by This could be an interesting field for further many factors, including poor water quality and at- experimental work and theoretical studies. tack by predators. It is deleterious to fish health on a number of levels: it can increase respiration rate and heart beat, which leads to energetic losses, ACKNOWLEDGEMENTS and can leave them more susceptible to disease (Helfman et al. 1997). While in hiding fish benefit We are grateful to the Ecology and Evolution 294 Chapter 13

Group at Leeds University for stimulating discus- Eklöv, P. and Persson, L. (1996) The response of prey to sions. J.K. was supported by the Fisheries Society the risk of predation: proximate cues for refuging juve- of the British Isles and the Leverhulme Trust. nile fish. Animal Behaviour 51, 105–15. Endler, J.A. (1986) Defense against predators. In: M.E. G.D.R. acknowledges support from the Nuffield Feder and G.V. Lauder (eds) Predator–Prey Relation- Foundation. E.M.A.H. was funded by a BBSRC ships. Chicago: University of Chicago Press, pp. studentship. 109–34. Endler, J. (1991) Interactions between predators and prey. In: J.R. Krebs and N.B. Davies (eds) Behavioural Ecology. Oxford: Blackwell, pp. 169–96. REFERENCES Ferguson, M.M. and Noakes, L.G. (1981) Social grouping and genetic variation in common shiners, N. cornutus Beukema, J.J. (1968) Predation by the three-spined stick- (Pisces, Cyprinidae). Environmental Biology of Fishes leback (Gasterosteus aculeatus L.). Behaviour 31, 6, 357–60. 1–126. Frank, K.T. and Leggett, W.C. (1985) Reciprocal oscilla- Blouw, D.M. (1996) Evolution of offspring desertion in a tions in densities of larval fish and potential predators: stickleback fish. Ecoscience 3, 18–24. a reflection of present or past predation? Canadian Brainerd, E.L. (1994) Pufferfish inflation: functional mor- Journal of Fisheries and Aquatic Sciences 42, 1841– phology of postcranial structures in Diodon holocan- 9. thus (Tetradontiformes). Journal of Morphology 200, Fraser, N.H.C. and Metcalfe, N.B. (1997) The costs of 1–20. becoming nocturnal: feeding efficiency in relation to Breder, C.M. (1946) An analysis of the deceptive resem- light intensity in juvenile Atlantic salmon. Functional blances of fishes to plant parts, with critical remarks Ecology 11, 385–91. on protective mimicry and adaptation. Bulletin of the Fraser, N.H.C., Metcalfe, N.B. and Thorpe, J.E. (1993) Bingham Oceanographic Collection 10, 1–49. Temperature-dependent switch between diurnal and Bremset, G. and Berg, O.K. (1999) Three-dimensional nocturnal foraging in salmon. Proceedings of the Royal microhabitat use by young pool-dwelling Atlantic Society of London Series B 252, 132–9. salmon and brown trout. Animal Behaviour 58, Giles, N. (1983) Behavioural effects of the parasite Schis- 1047–59. tocephalus solidus (Cestoda) on an intermediate host, Brönmark, C. and Miner, J.G. (1992) Predator-induced the three-spined stickleback, Gasterosteus aculeatus phenotypical change in body morphology in crucian L. Animal Behaviour 31, 1192–4. carp. Science 258, 1348–50. Godin, J.-G.J. (1986) Antipredator functions of shoaling Brown, R.S. (1999) Fall and early winter movements of in teleost fishes: a selective review. Naturaliste Cana- cutthroat trout, Oncorhynchus clarki, in relation to dian. 113, 241–50. water temperature and ice conditions in Dutch Creek, Godin, J.-J.G. (1997) Evading predators. In: J.-G.J. Godin Alberta. Environmental Biology of Fishes 55, 359–68. (ed.) Behavioural Ecology of Teleost Fishes. Oxford: Bumann, D., Krause, J. and Rubenstein, D.I. (1997) Oxford University Press, pp. 191–236. Mortality risk of spatial positions in animal groups: Greenwood, M.F.D. and Metcalfe, N.B. (1998) Minnows the danger of being in the front. Behaviour 134, 1063– become nocturnal at low temperatures. Journal of Fish 76. Biology 53, 25–32. Dafni, J. and Diamant, A. (1984) School-oriented mimi- Hamilton, W.D. (1971) Geometry for the selfish herd. cry, a new type of mimicry in fishes. Marine Ecology Journal of Theoretical Biology 31, 295–311. Progress Series 20, 45–50. Helfman, G.S., Collette, B.B. and Facey, D.E. (1997) The Dill, L.M. and Fraser, A.H.G. (1997) The worm Diversity of Fishes. Oxford: Blackwell Science. re-turns: hiding behaviour of a tube-dwelling poly- Hoogland, R.D., Morris, D. and Tinbergen, N. (1957) The chaete, Serpula vermicularis. Behavioral Ecology 8, spines of sticklebacks (Gasterosteus and Pygosteus) 186–93. as means of defence against predators (Perca and Esox). Dugatkin, L.A. and Godin, J.-G.J. (1992) Prey approach- Behaviour 10, 205–37. ing predators: a cost-benefit perspective. Annales Huhta, A., Muotka, T., Juntunen, A. and Yrjönen, M. Zoologie Fennici 29, 233–52. (1999) Behavioural interactions in stream food webs: Ehrlich, P. and Ehrlich, A. (1973) Coevolution: heterotyp- the case of drift-feeding fish, predatory invertebrates ic schooling in Caribbean reef fishes. American Natu- and grazing mayfly. Journal of Animal Ecology 68, ralist 107, 157–60. 917–27. Fish as Prey 295

Humphries, S., Ruxton, G. and Metcalfe, N.B. (1999) Licht, T. (1989) Discriminating between hungry and Patch choice and risk: relative competitive ability is satiated predators: the response of guppies (Poecilia context dependent. Animal Behaviour 58, 1131–8. reticulata) from high and low predation sites. Ethology Irving, P.W. and Magurran, A.E. (1997) Context- 82, 238–43. dependent fright reactions in captive European Losey, G.S. (1972) Predation protection in the poison- minnows: the importance of naturalness in laboratory fang blenny, Meiacanthus atrodorsalis, and its mim- experiments. Animal Behaviour 53, 1193–201. ics, Escenius bicolor and Runula laudadus (Bleniidae). Itô, Y. (1998) Role of escape from predators in periodical Pacific Science 26, 129–39. cicada (Homoptera: Cicadidae) cycles. Annals of the Magurran, A.E. (1990) The adaptive significance of Entomological Society of America 91, 493–6. schooling as an anti-predator defence in fish. Annales Karplus, I. and Algom, D. (1981) Visual cues for predator Zoologie Fennici 27, 51–66. face recognition by reef fishes. Zeitschrift für Tierpsy- Magurran, A.E., Oulton, W. and Pitcher, T.J. (1985) Vigi- chologie 55, 343–64. lant behaviour and shoal size in minnows. Zeitschrift Krakauer, D.C. (1995) Groups confuse predators by für Tierpsychologie 67, 167–78. exploiting perceptual bottlenecks: a connectionist Magurran, A.E., Irving, P.W. and Henderson, P.A. (1996) model of the confusion effect. Behavioral Ecology and Is there a fish alarm pheromone? A wild study and Sociobiology 36, 421–9. critique. Proceedings of the Royal Society of London Krause, J. (1993) The effect of ‘Schreckstoff’ on the Series B 263, 1551–6. shoaling behaviour of the minnow: a test of Hamilton’s Major, P.F. (1978) Predator–prey interactions in two selfish herd theory. Animal Behaviour 45, 1019–24. schooling fishes, Caranx ignobilis and Stolephorus Krause, J. (1994) Differential fitness returns in relation purpureus. Animal Behaviour 26, 760–77. to spatial positions in groups. Biological Reviews 69, Mathis, A., Chivers, D.P. and Smith, R.J.F. (1995) 187–206. Chemical alarm signals: predator-deterrents or pre- Krause, J. and Godin J.-G.J. (1994) Shoal choice in dator attractants? American Naturalist 145, 994– banded killifish (Fundulus diaphanus, Teleostei, 1005. Cyprinodontidae): the effects of predation risk, fish McCosker, J.E. (1977) Fright posture of the plesiopid size, species composition and size of shoals. Ethology fish Calloplesiops altivelis: an example of Batesian 98, 128–36. mimicry. Science 147, 400–1. Krause, J. and Godin, J.-G.J. (1995) Predator preferences McIntosh, A.R. and Peckarsky, B.L. (1996) Differential for attacking particular group sizes: consequences behavioural responses of mayflies from streams with for predator hunting success and prey predation risk. and without fish to trout odour. Freshwater Biology 35, Animal Behaviour 50, 465–73. 141–8. Krause, J., Godin, J.-G.J. and Brown, D. (1996) Phenotypic McKaye, K.R. and Oliver, M.K. (1980) Geometry of a variability within and between fish shoals. Ecology 77, selfish school: defence of cichlid young by a bagrid 1586–91. catfish in Lake Malawi, Africa. Animal Behaviour Krause, J., Ruxton, G.D. and Rubenstein, D.I. (1998a) Is 28, 1287. there an influence of group size on predator hunting Metcalfe, N.B., Fraser, N.H.C. and Burns, M.D. (1999) success? Journal of Fish Biology 52, 494–501. Food availability and the nocturnal vs. diurnal foraging Krause, J., Loader, S.P., McDermott, J. and Ruxton, G.D. trade-off in juvenile salmon. Journal of Animal Eco- (1998b) Refuge use by fish as a function of body length- logy 68, 371–81. related metabolic expenditure and predation risks. Milinski, M. (1977) Experiments on the selection by Proceedings of the Royal Society of London Series B predators against spatial oddity of their prey. 265, 2373–9. Zeitschrift für Tierpsychologie 43, 760–77. Krause, J., Loader, S.P., Kirkman, E. and Ruxton, G.D. Milinski, M. (1990) Parasites and host decision making. (1999) Refuge use by fish as a function of body weight In: C.J. Barnard and J.M. Behnke (eds) Parasitism and changes. Acta Ethologica 2, 29–34. Host Behaviour. London: Taylor and Francis, pp. Krause, J., Butlin, R., Peuhkuri, N. and Pritchard, V. 95–116. (2000) The social organisation of fish shoals: a test of Milinski, M.M. (1993) Predation risk and feeding the predictive power of laboratory experiments for the behaviour. In: T.J. Pitcher (ed.) The Behaviour of Teleost field. Biological Reviews 75, 477–501. Fishes, 2nd edn. London: Croom Helm, pp. 285–305. Landeau, L. and Terborgh, J.W. (1986) Oddity and the Moodie, G.E.E. (1972) Predation, natural selection and ‘confusion effect’ in predation. Animal Behaviour 24, adaptation in an unusual threespine stickleback. 1373–80. Heredity 28, 155–67. 296 Chapter 13

Naish, K.A., Carvalho, G.R. and Pitcher, T.J. (1993) The Randall, J.E. and Randall, H.A. (1960) Examples of genetic structure and microdistribution of shoals of mimicry and protective resemblance in tropical Phoxinus phoxinus, the European minnow. Journal of marine fishes. Bulletin of Marine Science of the Gulf Fish Biology 43 (Suppl. A), 75–89. and Caribbean 10, 444–80. Naud, M. and Magnan, P. (1988) Diel onshore–offshore Ranta, E., Lindström, K. and Peuhkuri, N. (1992) Size migrations in northern redbelly dace, Phoxinus eos matters when three-spined sticklebacks go to school. (Cope), in relation to prey distribution in a small Animal Behaviour 43, 160–2. oligotrophic lake. Canadian Journal of Zoology 66, Reimchen, T.E. (1994) Predators and morphological 1249–53. evolution in threespine stickleback. In: M.A. Bell and Neill, S.R.St.J. and Cullen, J.M. (1974) Experiments on S.A. Foster (eds) The Evolutionary Biology of the whether schooling by their prey affects the hunting Threespine Stickleback. Oxford: Oxford Science behaviour of cephalopod and fish predators. Journal of Publications, pp. 240–76. Zoology London 172, 549–69. Ritland, D.B. (1995) Comparative unpalatability of Nilsson, P.A., Brsˇnmark, C. and Pettersson, L.B. (1995) mimetic viceroy butterflies (Limenitis archippus) Benefits of a predator-induced morphology in crucian from 4 South-Eastern United-States populations. carp. Oecologia 104, 291–6. Oecologia 103, 327–36. Nottestad, L. (1999) Spatial aspects of life-cycle adapta- Rodewald, A.D. and Foster, S.A. (1998) Effects of gravi- tion and survival in adult herring. PhD thesis, Univer- dity on habitat use and antipredator behaviour in sity of Bergen, Bergen. three-spined sticklebacks. Journal of Fish Biology 52, Pallini, A., Janssen, A. and Sabelis, M.W. (1998) Predators 973–84. induce interspecific herbivore competition for food in Sargent, R.C. (1985) Territoriality and reproductive refuge space. Ecology Letters 1, 171–7. tradeoffs in the threespine stickleback, Gasterosteus Parker, G.A. (1984) Evolutionary stable strategy. In: J.R. aculeatus. Behaviour 93, 217–26. Krebs and N.B. Davies (eds) Behavioural Ecology: An Seigel, J.A. and Adamson, T.A. (1983) Batesian mimicry Evolutionary Approach. Oxford: Blackwell Scientific between a cardinalfish (Apogonidae) and a venomous Publications, pp. 30–61. scorpionfish (Scorpaenidae) from the Philippine- Parrish, J.K. (1989) Re-examining the selfish herd: are Islands. Pacific Science 37, 75–9. central fish safer? Animal Behaviour 38, 1048–53. Sih, A. (1992) Prey uncertainty and the balancing of Parrish, J.K., Strand, S.W. and Lott, J.L. (1989) Predation antipredator and feeding needs. American Naturalist on a school of flatiron herring, Harengula thrissina. 139, 1052–69. Copeia 1989, 1089–91. Sih, A. (1997) To hide or not to hide? Refuge use in a fluc- Perrin, N. (1995) Signalling, mating success and paternal tuating environment. Trends in Ecology and Evolution investment in sticklebacks (Gasterosteus aculeatus): 12, 375–6. a theoretical model. Behaviour 132, 1037–57. Smith, R.J.F. (1997) Avoiding and deterring predators. Peuhkuri, N., Ranta, E. and Seppä, P. (1997) Size- In: J.-G.J. Godin (ed.) Behavioural Ecology of Teleost assortative schooling in free-ranging sticklebacks. Fishes. Oxford: Oxford University Press, pp. 163–90. Ethology 103, 318–24. Sogard, S.M. (1997) Size selective mortality in the juve- Pitcher, T.J. and Parrish, J.K. (1993) Functions of shoaling nile stage of teleost fishes: a review. Bulletin of Marine behaviour in teleosts. In: T.J. Pitcher (ed.) Behaviour Science 60, 1129–57. of Teleost Fishes. London: Chapman & Hall, pp. 363– Sogard, S.M. and Olla, B.L. (1993) Effects of light, thermo- 439. cline and predator presence on vertical distribution Poyard, L., Desmarais, E., Chenuil, A., Agnese, J.F. and and behavioural interactions of juvenile walleye Bonhomme, F. (1999) Kin cohesiveness and possible in- pollock, Theragra chalcogramma Pallas. Journal of breeding in the mouthbrooding tilapia Sarotherodon Experimental Marine Biology and Ecology 167, 179– melanotheron. Molecular Ecology 8, 803–12. 95. Pulliam, H.R. and Caraco, T. (1984) Living in groups: is Svensson, P.A., Barber, I. and Forsgren, E. (2000) Shoaling there an optimal group size? In: J.R. Krebs and N.B. behaviour of the two-spotted goby. Journal of Fish Biol- Davies (eds) Behavioural Ecology: An Evolutionary ogy 56, 1477–87. Approach. Oxford: Blackwell Scientific Publications, Theodorakis, C.W. (1989) Size segregation and the effects pp. 122–47. of oddity on predation risk in minnow schools. Animal Randall, J.E. and Emery, A.R. (1971) On the resemblance Behaviour 38, 496–502. of the young of the fishes Platax pinnatus and Plec- Turner, G.F. and Pitcher, T.J. (1986) Attack abatement: a torhynchus chaetodontoides to flatworms and nudi- model for group protection by combined avoidance and branchs. Zoologica 197, 115–19. dilution. American Naturalist 128, 228–40. Fish as Prey 297 von Frisch, K. (1941) Über einen Schreckstoff der Winemiller, K.O. (1990) Caudal eyespots as deterrents Fischhaut und seine biologische Bedeutung. against fin predation in the neotropical cichlid As- Zeitschrift für Vergleichende Physiologie 29, 46–146. tronotus ocellatus. Copeia 1990, 665–73. Wainwright, P.C., Turingan, R.G. and Brainerd, E.L. Wootton, R.J. (1976) The Biology of the Sticklebacks. (1995) Functional morphology of pufferfish inflation: London: Academic Press. Mechanism of the buccal pump. Copeia 1995, 614–25. Part 4 Fish in Ecosystems 14 Trophic Ecology and the Structure of Marine Food Webs

NICHOLAS V.C. POLUNIN AND J.K. PINNEGAR

14.1 INTRODUCTION for such research or management (Botsford et al. 1997). A corollary of the sustainable use of marine re- One major scientific context for understand- sources is that these resources are not exploited in ing interactions among organisms is that of the isolation and a broad perspective is therefore desir- ecosystem, and ‘ecosystem management’ has be- able. The exploitation of any resource inevitably come a useful focal point among those probing the has wider implications (Christensen et al. 1996), scientific bases for a holistic and sustainable ap- and one of the most frequently expressed concerns proach in the use of natural resources (Haeuber and about intensive fishing is that it will lead to imbal- Franklin 1996). However, delving into the struc- ances in ecosystem function that have ramifica- ture and workings of ecosystems soon reveals the tions for community structure overall (Walters piecemeal and fragmentary nature of the informa- and Holling 1990; Jennings and Kaiser 1998; Pauly tion available (McCann et al. 1998), yet such en- and Christensen, Chapter 10 and Kaiser and quiry is essential if the science of management is Jennings, Chapter 16, Volume 2). The changes in- to develop and is to address more appropriately the volved need not always be ‘negative’; for example, purpose of sustaining the supply of environmental it is thought that removal of large predators will goods, i.e. the fish and services such as waste sometimes increase the abundance and productiv- removal, in coastal waters. ity of prey organisms and these may benefit fish- An important class of interactions which oc- eries (Jennings and Kaiser 1998; Pauly et al. 1998a). curs within ecosystems is that involving fluxes of However in other cases, the removal of predators nutrients and energy, the study of which is gener- by fishing has resulted in increases in the abun- ally referred to as trophodynamics. Food chains dance of organisms such as sea urchins, which at and webs specifically aim to describe the trophic high densities often have detrimental effects on relationships amongst organisms in a community communities as a whole (review in Pinnegar et al. and are a major focus of trophodynamic studies. An 2000). There is an emerging consensus among fish- appealing feature of trophic fluxes is that they can eries scientists and managers of aquatic resources be quantified in common currencies, such as units that conventional single-species analyses ought to of energy, mass or carbon. They link many of the be replaced by approaches that explicitly account components of ecosystems and have long been for ecological interactions, and especially those viewed by some as the cornerstone of ecosystem of a trophic nature (Walters et al. 1997; Pauly and ecology (Odum 1968). Such concepts apparently Christensen, Chapter 10, Volume 2). However, underpin the notion that each human action, how- there is little consensus concerning the analytical ever specialized and local, inevitably has wider and conceptual tools that should be used as a basis consequences, and these may at times include sur- 302 Chapter 14

Open ocean Coastal zone Upwelling

6 Piscivores

5 Planktivores

4 Megazooplankton Piscivores

Trophic level 3 Macrozooplankton Planktivores Benthic Megazooplanktivores carnivores

2 Microzooplankton Macrozooplankton Benthic Megazooplankton Planktivores herbivores

1 Nanoplankton Microphytoplankton Macrophytoplankton and nanoplankton

Fig. 14.1 Food chains postulated by Ryther to characterize the three marine provinces that he identified as a basis for estimating global marine fish productivity; the arrows indicate fluxes of organic carbon to consumers. (Source: after Ryther 1969.) prises and failures that influence the sustainability this and (iv) evaluate the roles of fish in food webs of the activities involved. Yet great uncertainties and consequences of changes in their abundance result from inferring secondary effects within as a result of management controls or exploitation. ecosystems from trophodynamic linkages alone, Aspects of individual trophic behaviour are and Menge (1992, 1997) has provided a detailed pic- analysed in Chapter 11–13 of this volume. The ture of the relative strengths of direct and indirect application of an ecosystem perspective to fish- interactions within particular marine systems eries management is presented by Pauly and based on small-scale experiments. Christensen in Chapter 10, Volume 2. The aim of the present chapter is to explore the origins and possible implications of the uncer- tainty in trophic interactions and to consider how 14.2 FOOD CHAINS AND information on indirect effects of exploitation and FOOD WEBS management might be derived at scales appropri- ate to an analysis of the possible effects of fishing As with much other information on ecological or the consequences of creating marine protected communities, there are two possible extremes in areas (discussed explicitly by Polunin in Chapter arranging feeding relationships on the basis of the 14, Volume 2). The specific objectives of the chap- available data. One of these extremes is to combine ter are to (i) consider limitations of the use of food species into groups that are trophically similar chains to predict processes such as fish production, (Fig. 14.1), while the other is to represent all rela- (ii) assess current understanding of food webs and tionships that exist among species in a complex what this is based on, (iii) assess variations in inter- reticulated system (Fig. 14.2). The first of these, the action strength in food webs and implications of food chain, may be used because data are in short Marine Food Webs 303

Young herring Adult 7–12 mm 12–42 mm 42–130 mm Herring

Tomopteris Pleuro Sagitta Limacina Oikopleura Medusae brachia

Ammodytes juv Decapod Nyctiphanes Cypris larvae Hyperiid balanus amphipods Larval larvae mollusca Evadne Pseudo- Acartia Temora Calanus Podon calanus Tintinnopsis

Peridinium Diatoms and flagellates (plants)

Fig. 14.2 Representation of the planktonic food web supporting the Atlantic herring, Clupea harengus, and including predators of the larvae (broken arrows); arrows indicate feeding by predators on prey organisms. (Source: from Hardy 1959. Reproduced by permission.) supply and separate relationships between species linear chain, between which fluxes and transfer cannot reliably be depicted, or it might be that the efficiencies can be traced. The apparent simplicity data do exist but by aggregation the dominant and elegance of the food-chain concept and the interactions are highlighted. The second extreme, ability to quantitatively compare systems in terms the food web, supposes that detailed information is of two variables, namely trophic level (an integer) available and that meaningful patterns and proper- and the efficiency of food-chain transfer (expressed ties can be gleaned from such a complex arran- as a fraction), has proved overwhelmingly attrac- gement. While the first approach predates the tive to many ecologists (studies reviewed in second, it has nevertheless persisted in various Cushing 1975). forms (Ulanowicz 1995; Jennings et al. 2001, 2002) Ryther (1969) used simple food-chain theory to and the purpose here is to consider the range of predict marine fishery productivity of the world’s possible approaches, beginning with their his- oceans. To do this he divided the oceans into three torical development and reviewing their relative ‘provinces’, the open ocean, the coastal zone and strengths and weaknesses. upwelling areas. These provinces vary substan- tially in primary productivity and, in addition, the 14.2.1 Food chains size of the most prevalent primary producers also differs significantly. It was supposed that the larger Food chains owe much to the early work of Elton the plant cells at the base of the food chain, as often (1927) and Lindeman (1942). Typically they rely found in nutrient-rich areas, the smaller the num- upon the aggregation of functionally similar ber of trophic levels needed to convert this organic species into groups or trophic levels arranged in a matter into a form useful for fishery exploitation. 304 Chapter 14

Table 14.1 Ryther’s (1969) estimates of global marine fish productivity based on food-chain characterization of production and assuming the existence of distinct ‘provinces’ of contrasting productivity, together with correct knowledge of their areas, net primary productivities, numbers of trophic levels and mean food-chain efficiencies.

Province Area Mean net primary Total net primary Trophic levels Efficiency Fish productivity (106 km2)productivity productivity (%) (fresh 10 6 tonnes year-1) (gm-2 year-1)(10 9 tonnes year-1)

Open ocean 326 50 16.3 5 10 0.1 Coastal zone 36 100 3.6 3 15 120 Upwelling areas 0.36 300 0.1 1.5 20 120

For example, the nanoplankton predominant in that they were likely to be fully exploited within open oceans cannot be captured directly by most the next decade or so (Cushing 1975). More re- crustacean zooplankters such as copepods, requir- cently Pauly and Christensen (1995) calculated the ing additional trophic levels consisting of interme- primary productivity required to sustain global diate-sized microzooplankton, such as protozoans fisheries catches, on the basis of a mean energy and larval nauplii of microcrustaceans, that are transfer efficiency between trophic levels of 10%, then consumed by copepods, and so on (Fig. 14.1). based on 48 Ecopath models of aquatic ecosystems Ryther (1969) further supposed that the least pro- (Pauly and Christensen, Chapter 10, Volume 2) and ductive oceanic regions had the lowest efficiencies fractional estimates of trophic level as opposed to of carbon transfer along food chains. Where food is the discrete integers used by Ryther (1969). It was less available animals must perform more work to estimated that around 8% of global aquatic pri- hunt, locate and capture prey, and thus less of the mary production was required to support annual food consumed is converted to production. world catches of 94 million tonnes plus 27 million Combining efficiency estimates with the tonnes of discarded bycatch; in shelf and fresh- number of trophic levels, Ryther (1969) was able to water systems as much as 24–35% may be re- contrast the high fish productivity of upwelling quired. Estimating transfer efficiencies between regions, which form only 0.1% of oceanic waters, trophic levels was previously a matter of much with the extremely low fish productivity of the conjecture and usually pertained to single-species open ocean systems, which make up 90% of populations (Slobodkin 1961). The values obtained oceanic waters (Table 14.1). Overall fish produc- by Pauly and Christensen (1995), though unsur- tivity of the world was estimated to be approxi- prising in terms of their mean (Morowitz 1991), mately 240 million tonnes annually (Table 14.1). differed radically from Ryther’s (1969) guesses at Assuming that only half or so of this productivity the mean transfer efficiencies of different pro- might be available sustainably as yield, Ryther vinces (Pauly et al. 2000). (1969) tentatively suggested a potential total The weaknesses of the approach adopted by marine fishery harvest of 100 million tonnes an- Ryther (1969) and subsequent authors, and the nually. Similar annual estimates (102–142 million food-chain approach generally, are many (Alverson tonnes) were obtained by Gulland (1970) using et al. 1970; Lasker 1988); for example, predictions more extensive primary productivity and transfer of the levels of biomass yield for different regions efficiency data, and these can be considered close of the world ocean are open to disagreement. to the current world catch as estimated from Nevertheless, with analyses of the critical state of recorded landings (FAO 1995). The most impor- world fisheries becoming ever more abundant, the tant consequence of this early practical use of food- shortcomings of Ryther’s study provide a number chain studies was the discovery that fish resources of points to discuss. General issues highlighted by in the world’s oceans are potentially limited and Ryther’s (1969) analysis include the following. Marine Food Webs 305

1 Feeding relationships among organisms in tivity data from the NASA Coastal Zone Colour ecosystems are more appropriately represented Scanner (CZCS) and 21872 chlorophyll depth pro- as food webs rather than food chains, since many files measured in the ocean. This system, which species, including those targeted by fisheries, feed is reviewed in detail in Longhurst (1998), consists at more than one trophic level (Polis and Strong of four primary ‘domains’, namely polar, wester- 1996). lies (temperate), trade-winds (tropical) and coastal, 2 The food-chain concept implies that size and/or and these are themselves partitioned into 57 sec- productivity of consumer populations are driven ondary biogeochemical provinces. These range in by primary productivity (i.e. ‘bottom-up’), al- size from the Red Sea and Persian Gulf province though other processes in food webs such as preda- (0.56 million km2) to the South Pacific Subtropical tion may also be very influential. Gyre (37.29 million km2), and in productivity 3 Ecological efficiencies will vary among types from the North Pacific Tropical Gyre (59gm-2 of organisms and among developmental stages year-1) to the Northeast Atlantic Continental within species (e.g. Pimm 1988). Thus aggregate Shelf (730gm-2 year-1). The total estimated annual efficiencies for trophic levels are likely to be primary productivity of the global oceans was unrealistic. 45–50 ¥ 109 tonnes, which is more than double that 4 Materials egested or not used by the organisms estimated by Ryther (1969) (Table 14.1) and sig- in food chains are not necessarily lost from the nificantly greater than the value of Pauly and ecosystems involved; for example, detritus may be Christensen (1995) (43.9 ¥ 109 tonnes). used by detritivores, with important implications for fish productivity (Polis and Strong 1996). 14.2.2 Food webs Indeed, much of the dissolved and particulate organic matter released by marine consumers is Food webs are representations of the feeding rela- now thought to return to the food web via the tionships within whole communities of organ- ‘microbial loop’ (Azam et al. 1983), which largely isms based principally on data obtained through consists of bacteria, heterotrophic flagellates and dietary analysis (Fig. 14.2). Such analyses typically microzooplankton (detritus recycling is also given involve gut contents data, and a wealth of different prominence by Pauly and Christensen in Chapter methods have been developed for use in studies of 10, Volume 2). fish (Hyslop 1980; Jobling, Chapter 5, this volume). 5 Ryther (1969) predicted that oceanic upwelling From the North Sea, in particular, stomach con- areas would have shorter food chains than the tents data have been extensively collected (Daan oligotrophic open oceans (Fig. 14.1). Energy is 1989) for use in the construction of multispecies cumulatively lost through processes such as defe- fisheries models (Magnusson 1995; Shepherd cation and excretion as a food chain is ascended and Pope, Chapter 7, Volume 2). However, gut and thus a high-level carnivore should require a content analyses have a number of limitations substantial home range in order to sustain it. Sites (Deb 1997); for example, they tend to provide mere with high primary productivities should exhibit snapshots of diets at particular points in time and longer food chains than those with low productiv- space. Moreover, the durations of the snapshots ities, since more energy is available for transfer vary amongst species, given variations in gut upwards to support predators at increasing trophic turnover (Jobling, Chapter 5, this volume), and levels (Pimm 1982). neglect certain types of dietary materials, such as It is not clear how ecologically appropriate it gelatinous plankton and detritus, that may never- was to divide up the world’s oceans into the three theless be very important. The significance of ‘provinces’ of Ryther (1969) or the five of Pauly this point is emphasized by Righton et al. (2001) and Christensen (1995). Longhurst et al. (1995) who show that feeding behaviour of cod (Gadus proposed a system for dividing up the ocean into morhua) varies temporally in the North Sea biologically coherent zones using primary produc- and Irish Sea. Organisms at the base of food webs 306 Chapter 14 tend to be particularly poorly characterized (Hall primary materials such as higher plants, benthic and Raffaelli 1993) and excessive lumping together algae and phytoplankton tend to differ in the rela- of species may greatly affect the apparent proper- tive abundance of carbon stable isotopes (Harrigan ties of food webs (Goldwasser and Roughgarden et al. 1989), while the nitrogen stable isotope data 1997). Gut contents therefore offer a poor basis for are good indicators of trophic level (Vander Zanden elucidating the detailed interactions which must et al. 1997; Kline and Pauly 1998). Such finger- exist (Paine 1988). Immunological assays of gut prints are likely to show a number of characteris- contents may be useful in such circumstances to tics in common. For example, when two widely supplement visual data on diets (van der Veer et al. separated coral reef sites in the South Pacific 1998), although in general most links in food (Fig. 14.3) are compared, piscivorous fishes exhibit webs remain unsuccessfully identified (Cohen and the highest values of the nitrogen stable isotope Newman 1988). index d15N, meaning that they are enriched in There are additional tools available for elucidat- the isotope 15N, benthic algae exhibit a higher ing food-web structure. A principal method in- d13C index, which means they are enriched in volves the use of stable isotopes, in particular the isotope 13C compared with planktonic materi- those of carbon and nitrogen (Pinnegar and als, and planktivorous and herbivorous fish have Polunin 2000). Carbon and nitrogen each exist in mean d15N values greater than the plankton and two stable isotopic forms (12C and 13C, 14N and algae they consume (Fig. 14.3). In addition to this, 15N) that differ in atomic mass. The least abundant piscivores typically exhibit somewhat intermedi- isotopes (13C and 15N) contain one more neutron ate d13C values between benthic-algal and plank- and are thus one atomic mass unit heavier than the tonic sources (Fig. 14.3), suggesting that both most abundant isotopes (12C and 14N). Chemical sources may ultimately contribute to production fractionation occurs because a chemical bond that of these fishes. Many of the similarities between involves a heavier isotope has a lower vibrational the Pacific reef sites are also evinced in the isotopic frequency and is therefore stronger than an equiva- profiles of Caribbean coral reefs (Fig. 14.4a) and lent bond involving the lighter isotope. Thus the Mediterranean temperate rocky reefs (Fig. 14.4b). probability of a bond breaking is greater for com- Thus similar isotopic patterns can be discerned pounds containing 14N or 12C than for those con- in the food webs of widely separated systems, taining 15N or 13C. When an animal consumes its highlighting the fact that these data reflect repeat- food, the unreacted substrate (i.e. faeces) becomes able and meaningful features of the food webs enriched in the heavier isotope compared with the concerned. reacted product, which is absorbed into the body. Other uses for stable isotopes include the eluci- However, the processes of catabolism, transami- dation of migration patterns (Hobson 1999), since nation and deamination and the resulting excre- organisms moving between isotopically distinct tion of products such as carbon dioxide, ammonia food webs can carry with them information on the or urea during respiration or assimilation result in location of previous feeding. Such an approach has the animal tending to become enriched rather than proven particularly useful to track inshore versus depleted in the heavier isotope with respect to offshore and marine versus freshwater movements their food. The 15N concentration in tissues of in anadromous fishes (Hesslein et al. 1991). Stable consumers is typically enriched by ~3‰ d units isotopes of nitrogen and carbon can reveal ontoge- relative to those of their prey (Minagawa and netic dietary changes in both fish (Lindsay et al. Wada 1984), whilst tissues are enriched in 13C by 1998) and invertebrates (Hentschel 1998) and have only ~1‰ as trophic level increases (DeNiro and been used to examine whether such changes occur Epstein 1978). earlier or later under differing fishing regimes When major component organisms of a commu- (Persson and Hansson 1999). Several studies, nity are analysed, the resulting data can constitute notably those of Wainright et al. (1993) and a form of ‘fingerprint’ of the food web because Thompson et al. (1995), have used stable isotope Marine Food Webs 307

(a) 14 (b) 14 F 12 12 C1 10 F1 F2 F3 P1 10 P1 P2 P1 H2 8 O1 H3 8 H1

N O1 C2 N 15 H2 15 H3 δ 6 δ 6 02 H1 H4 C1 H4 Z1 Z2 4 A2 4 A1 A1 A3 A4 2 2 A2 O2 0 0 –20 –18 –16 –14 –12 –10 –8 –6 –4 –2 –20 –18 –16 –14 –12 –10 –8 –6 –4 –2 δ13C δ13C

Fig. 14.3 Comparison of the isotopic ‘fingerprints’ of two shallow coral reef food webs, 3000km apart in the South Pacific, based on the natural abundance data (mean ± SE, where visible) of carbon (d13C) and nitrogen (d15N). (a) Northwestern outer margin of the Great Astrolabe Reef, Dravuni, Fiji. Algae: A1, algal turf in territories of the fish Plectroglyphidodon lacrymatus; A2, calcareous encrusting algae; A3, algal turf in territories of the fish Acanthurus lineatus; A4, algal turf in territories of the fish Stegastes nigricans. Coral-feeding fishes: C1, Chaetodon trifasciatus; C2, Chaetodon baronessa. Piscivorous fishes: F1, Aphareus furca; F2, Plectropomus leopardus; F3, P. laevis. Planktonic materials: O1, POM >53mm collected in March; O2, POM >53mm collected in September. Plankton- feeding fishes: P1, Pseudanthias pascalus; P2, Caesio caerulaureus. (b) Barrier reef at Tiahura, Moorea, French Polynesia. Algae: A1, algal turf in territories of the fish Stegastes nigricans; A2, calcareous encrusting algae. Coral-feeding fish: C1, Chaetodon trifasciatus. Piscivorous fish: F, Cephalopholis argus. Herbivorous fishes: H1, Acanthurus nigrofuscus; H2, Stegastes nigricans; H3, Ctenochaetus striatus; H4, Chlorurus sordidus. Planktonic materials: O1, POM >53mm collected during the day; O2, POM >53mm collected at night. Plankton-feeding fishes: P1, Heteroconger sp.; P2, Myripristis violaceus. Reef-building corals: Z1, Acropora sp.; Z2, Porites sp. (Source: N.V.C. Polunin, unpublished data.)

compositions in archival tissues of predatory fish tissue turnover is more rapid. Monteiro et al. and seabirds to look for long-term changes in the (1991) focused on muscle lipids because they trophic structure of North Atlantic food webs, and sought to highlight relatively short-term dietary have tentatively related such changes to changing histories of anchovies. On the other hand, removal patterns of exploitation over the past 50–100 years. of lipid is likely to reduce the variability of isotope It is becoming progressively less expensive to data (Pinnegar and Polunin 1999). measure stable isotopes in biological tissues, and Gut contents and stable isotope data each have technological advances are making the method their own relative merits (Table 14.2). The uses of increasingly available to ecologists. A major gut contents data are reviewed by Jobling in Chap- advantage of these data over conventional dietary ter 5, this volume. These data have strengths in re- analyses is that the tissues that are sampled turn solving individual trophic linkages, but the extent over relatively slowly and stable isotope composi- of these is limited to those parts of the diet that tions therefore reflect the diet of the animal over are readily identified and quantified. Stable isotope substantial periods of time. In the case of most late data can be especially useful for assessing the rela- juvenile and adult fishes, the time-scale concerned tive importance of major pathways (Doucett et will be of a year or more (Hesslein et al. 1993), but al. 1996), whether two species feed similarly or will tend to be shorter for very young fishes dissimilarly, and for indicating the component of (Doucett et al. 1996) and other organisms in which the diet that is actually assimilated and not simply 308 Chapter 14

(a) 14 (b) 14

12 12 BV SB MH 10 HE GC 10 DS OM CO AI LJ CR CC CF HR SE DD 8 EG N 8 SO HF 15 SP N AN SM DA CJ δ

15 HA δ PC ES P3 AP ST 6 OC 6 BB SS MS CB P4 B10 B5 SP CAC B8 P2 B9 B6 4 B4 4 B6 B1 B2 P1 B7 B1 B4 B5 A5 B7 B8 P5 A1 B3 2 P1 A1 2 B3 A3 B2 A4 A3 A2 A4 B1 A2 0 0 –20 –18 –16 –14 –12 –10 –8 –6 –4 –2 –24 –23 –22 –21 –20 –19 –18 –17–16 –15 –14 –13 δ13C δ13C

Fig. 14.4 Isotopic ‘fingerprints’ of (a) Caribbean coral reef and (b) Mediterranean rocky littoral community (Revellata Peninsula, Corsica). Plants are represented by hexagons, benthic invertebrates by squares, pelagic invertebrates by diamonds and fishes by circles. (a) Caribbean. Algae and seagrasses: A1, Batophera sp.; A2, Halimeda sp.; A3, Syringodium filiforme; A4, Thalassia testudinum. Invertebrates: B1, Bittium varium; B2, Chiton squamosus; B3, Diadema antillarium; B4, Grapsus sp.; B5, Panulirus argus; B6, Serpulidae; B7, Strombus gigas; B8, Tripneustes esculentus; P1, plankton. Fishes: AN, Aetobatus narinari; BV, Balistes vetula; CB, Calamus bajonado; CF, Cephalopholis fulva; CR, Carangoides ruber; EG, Epinephelus guttatus; ES, Epinephelus striatus; GC, Ginglymostoma cirratum; HA, Haemulon album; HE, Hemiramphus sp.; HF, Haemulon flavolineatum; HR, Holocentrus rufus; LJ, Lutjanus jocu; OC, Ocyurus chrysurus; PC, Priacanthus cruentatus; SB, Sphyraena barracuda; SP, Sparisoma sp.; Star, green turtle Chelonia mydas. (Source: data from Keegan and DeNiro 1988.) (b) Mediterranean. Algae and seagrasses: A1, Cystoseira balearica; A2, Cladophora prolifera; A3, Halopteris scoparia; A4, Dictyota dichotoma; A5, Posidonia oceanica. Benthic invertebrates: B1, Paracentrotus lividus; B2, Sphaerechinus granularis; B3, Haliotis lamellosa; B4, Cerithium vulgatum; B5, Bittium reticulatum; B6, Eunice harassi; B7, mixed benthic amphipods; B8, Leptomysis livingula; B9, Pagurus chevreuxii; B10, Lysmata seticaudata; B11, Xantho poressa. Pelagic invertebrates: P1, zooplankton >200mm; P2, Anchylomera blossevillei; P3, Siriella spp.; P4, Pelagia noctiluca; P5, Chelophyes appendiculata. Fishes: AI, Apogon imberbis; AP, Atherina presbyter; BB, Boops boops; CC, adult Chromis chromis; CCj, juvenile Chromis chromis; CJ, Coris julis; CO, Conger conger; DA, Diplodus annularis; DD, juvenile Dentex dentex; DS, Diplodus sargus; MH, Muraena helena; MS, Mullus surmuletus; OM, Oblada melanura; SE, Serranus scriba; SM, Spicara maena; SO, Symphodus ocellatus; ST, Symphodus tinca; SP, Scorpaena porcus. (Source: data from Pinnegar and Polunin 2000.) that which is ingested. In addition, stable isotope different tissues or body processes compared with analysis has proven useful for measuring trophic those derived from other sources, such as plant level (Minagawa and Wada 1984) and for discern- carbohydrates. The upshot of this is that tissues ing spatial variations in nutrition (Jennings et al. may not always reflect the isotopic composition 1997). There are also constraints to the use of of the bulk diet and might actually reflect the iso- stable isotopes and their interpretation (Table topic composition of an apparently minor dietary 14.2); one such constraint, relevant to the elucida- constituent. This phenomenon may be particu- tion of food webs, involves the phenomenon larly apparent in herbivorous animals such as the known as ‘isotopic routing’ (Gannes et al. 1997). It parrotfish Sparisoma spp. (Fig. 14.4a) and the would appear that the isotopes obtained in the diet sparid Sarpa salpa (Fig. 14.4b), which exhibit d15N of a particular consumer are not always evenly ap- values higher than would be expected if they were portioned once ingested, such that the nitrogen feeding on plant material. Since it is more efficient and carbon derived from certain dietary compo- to assimilate proteins and to catabolize carbohy- nents, such as animal proteins, may be routed to drates, and because the plant material consumed Marine Food Webs 309

Table 14.2 Advantages and disadvantages of gut contents and stable isotope analyses for elucidating the trophic relationships of consumers.

Information Gut contents Stable isotopes

Resolution of principal trophic Can be good where individual sources are Can be good if pathways well distinguished by d13C pathways in food webs identifiable (e.g. indigestible hard parts) of the basal materials, poor if more than two pathways Connectance (proportion of Good but only for individual sources that Poor because only broad categories distinguishable linkages that are realized) are identifiable as a rule Measure of nutritional roles of Poor because diet, not actual absorption, Can be good because isotopes are in materials that different dietary items quantified have been assimilated Measure of short-term differences Potentially good because data are only Poor because tissue turnover slow in diet of large predators short term Measure of spatial differences in Will be good where major items Will be good where shifts occur in items with diet of large predators identifiable distinct d13C and/or in trophic level Measure of trophic level Often inaccurate because diet Can be accurate if basal materials identified, and incompletely described change in d15N per trophic level validated

by herbivores largely consists of carbohydrates changes, differences in fatty acid composition are but is often supplemented with small amounts of detectable in the tissues of animals that have fed proteinaceous animal materials (Robertson 1982), on different diets. As is the case with stable iso- the contribution of plant sources in the diet of topes, these differences reflect dietary integration herbivorous fishes may be greatly underestimated over longer time periods (Sargent et al. 1988; Fraser by stable nitrogen isotope analyses. et al. 1989). In laboratory diet-switch experiments, Stable isotope data and conventional dietary tissue fatty acid patterns may change significantly data have thus together contributed to the elucida- in only 3 weeks in adult cod to reflect the patterns tion of food-web structure (Pinnegar and Polunin found in the new food (Kirsch et al. 1998). Thus gut 2000) and both methods have revealed features contents analysis indicates dietary patterns over that would not otherwise have been detected using short time periods of hours to days, stable isotopes the other method. can be used to elucidate patterns over longer time A further indirect method that can yield useful periods of months to years, and fatty acid analysis information about a consumer’s diet is the analy- may reflect diet over periods of intermediate dura- sis of fatty acids (Iverson 1993). When triglycerides tion of weeks to months. and other forms of lipid are ingested, they are bro- Because of their great variety, long-chain and ken down into free fatty acids and monoglycerides polyunsaturated fatty acids would appear to be before the body can incorporate them. Ingested particularly suitable as trophic tracers in marine fatty acids that are not utilized for immediate organisms. It has been demonstrated that specific energy requirements are generally stored without fatty acids or combinations of them are associated substantial modification in adipose tissue (Iverson with certain taxonomic classes of phytoplankton et al. 1995). Chain elongation and insertion of dou- (Sargent 1976, 1978). Thus species of Chryso- ble bonds may occur to some extent when fatty phyceae, Haptophyceae and Dinophyceae are acids are incorporated into cell membrane struc- characterized by the presence of C18:4 (n - 3) and ture (Hagen et al. 1995; Pond et al. 1997), and some C18:5 (n - 3) fatty acids, which are essentially biosynthesis can occur from other dietary compo- absent from diatoms. Diatoms are typified by the nents such as amino acids. Despite these potential fatty acids C16:4 and C20:5 (n - 3) as well as hav- 310 Chapter 14 ing a higher ratio of C16:1 (n - 7) to C16:0 fatty consumers that are well represented may also be acids than other phytoplankton classes, thus pro- poorly characterized because of a paucity of dietary viding the potential for identifying the transfer of data for them (Cortés 1999). different sources of production through a number 4 As with food chains, food webs indicate trophic of trophic levels. Analysis of fatty acid profiles fluxes among organisms in the system, but feeding have been widely used to elucidate the diets of fish, is only one of many types of interaction among or- marine mammals and seabirds (Iverson et al. ganisms. Processes such as migration and recruit- 1997), to differentiate between marine and fresh- ment may also be very influential in controlling water sources (Smith et al. 1996) and to highlight population sizes (Jones 1991) and thus food-web differences between geographically separated structures. populations of the same species (Iverson et al. In most cases the knowledge base of food webs 1997; Logan et al. 2000). They have also been used is very thin (Goldwasser and Roughgarden 1997), to look for seasonal and ontogenetic patterns in the but it is worth highlighting a few further character- diet of fish and invertebrate species (Pedersen istics of food webs before proceeding. Food webs et al. 1999), and different species of marine fish, vary greatly both spatially and temporally, at the cephalopods and crustaceans vary markedly in top especially through fluctuations in processes their fatty acid profiles (Iverson et al. 1997; Raclot such as recruitment and fishing, and at the bottom et al. 1998). by means of variations in primary productivity in- Complicated food-web diagrams such as that cluding the effects of seasonality. In some cases, illustrated in Fig. 14.5a have become the most concomitant changes may occur across all trophic common way of visualizing the complex interac- levels at large spatial scale and be linked to tions that exist within communities. However, long-term climatic trends (Aebischer et al. 1990; such depictions may have many shortcomings Polovina et al. 1994). The boundaries of food webs as a basis for predicting indirect impacts of may often be hard to define (Pimm et al. 1991; human interventions or natural changes in the Pauly and Christensen, Chapter 10, Volume 2), environment. Possible weaknesses include the given the extent of export and import of organic following. matter and of large-scale migrations of animals 1 Diets vary temporally and spatially (Righton (Hesslein et al. 1991). Omnivory is often wide- et al. 2001), yet most food-web studies assume spread and may reduce the potential for indirect steady-state conditions and data are derived from effects such as trophic cascades and ‘prey-release’ very short-term studies conducted at small spatial (Kaiser and Jennings, Chapter 16, Volume 2). This scales (Frid and Hall 1999). Composite food webs, is because predators in the community can feed where they exist, build upon data from many areas opportunistically at more than one trophic level and points in time and may offer a promise of gen- (Monteiro et al. 1991; Polis and Strong 1996). erality, but then beg questions of local accuracy In spite of evidence that certain trophic (Paine 1988). pathways may be relatively long and that there is 2 Many dietary items, such as dissolved organic a relationship between web size, or the level of matter, particulate detritus, gelatinous zooplank- aggregation, and mean food-chain length (Hall and ton, fish eggs and spat, are often poorly quantified. Raffaelli 1993), individual trophic pathways Thus certain pathways within systems, which within webs rarely consist of more than four may support a large part of fish production, will trophic levels (Hall and Raffaelli 1993; Pauly et al. scarcely be described (Hall and Raffaelli 1993). 1998b). Many factors may limit the length of food 3 Particular organisms, notably bacteria, fungi, chains (Pimm 1982, 1991; Sterner et al. 1997). protozoans, nematodes and other small meta- Amongst the most obvious is primary produc- zoans, typically tend to be greatly underrepres- tivity, yet food chains in productive systems are ented and/or incorrectly assigned in food webs, yet not necessarily shorter than those in oligotrophic may play important roles (Pimm 1982). Larger systems (Briand and Cohen 1987). Marine Food Webs 311

16.3 0.65 (a) 2.59 Mullet Stingray 1.30

0.07 8.15 0.51 2.14 1.57 0.01 404 24.4 0.03 1.26 Bay 652 Microphytes 11.2 1.57 0.16 anchovy 0.08 0.64 1.32 0.50 0.37 0.74 194 44.7 1.00 0.73 Gulf 1.50 0.16 73.1 Benthic 2.67 killfish invertebrate 0.17 2.68 0.14 feeders 0.14 0.0 Zooplankton 0.37 3.70 0.01 0.43 0.03 0.01 110 3.38 0.02 0.23 0.29 0.34 39.9 0.65 31.3 5.86 0.64 Silverside 7.23 0.64 Needlefish 0.97 0.06 0.36 1191 Detritus 1.49 1.53 0.12 8.50 0.09 568 3.20

0.94 0.06 2530 0.74 2120 687 0.90 Moharra 0.02 0.56 0.41

4160 2.30 Benthic 2.94 0.01 invertebrates 0.23 Pinfish 0.53 0.16 0.09 0.68 0.38 982 428 Goldspotted 0.10 1.16 0.71 killifish 0.01 0.07 6700 Macrophytes 0.15 0.32 0.65 0.89 1.06 Sheepshead 0.79 killifish 2320 219 Longnosed killifish 0.44 0.54 0.14 5.28 1.27 2.87

(b) 787 471 5.41 1.32 0.02

7360I + D 2311II 35III 5.58IV 0.065 V 28.4% 1.5% 16.9% 1.2% 0%

4570 5050 1030 16.5 3.070.03 778 7.48 1.13 0.013 787

Fig. 14.5 Carbon flows (mgm-2 day-1) among taxa of a marsh ecosystem, Crystal River Florida: (a) schematic food web; (b) aggregated chain of trophic transfers derived from (a). (a) Unfilled arrows (Æ) depict returns to detritus; diamond-shaped arrows ( ) represent respiration; sharp arrows ( ) depict import or export; triangular arrows ( ) represent trophic fluxes. (b) Flows out of the top of boxes represent export and flows out of the bottom represent respiration. Recycling of non-living material is through compartment D (detritus) and the percentages within the boxes represent annual trophic efficiencies. (Source: after Ulanowicz 1995.) 312 Chapter 14

Ulanowicz (1995) has provided procedures for within a food web, for example through distur- passing readily from complex food webs (Fig. bance of one of its components, can be expected 14.5a) to simple concatenated chains of trophic to be rapidly attenuated with each trophic step interactions (Fig. 14.5b). Such chains, Ulanowicz (Strong 1992). (1995) claimed, are useful because they provide A way forward in understanding functional an accurate but uncomplicated picture of the sys- linkages among organisms, or groups of organisms, tem’s underlying trophodynamics, while making will be found only where empirical data on interac- it possible to map arbitrarily complicated net- tion strengths are available. The next step is there- works into a common topological format. The fore to consider how relevant information of this method involves the construction of a matrix of sort can be derived. feeding coefficients and the allocation of the differ- ent dietary interactions, but not the consumer it- self, into discrete trophic levels (sensu Ulanowicz 14.3 INTERACTION 1986). Cycles within the system via detritus are STRENGTH IN FOOD WEBS treated separately (Ulanowicz 1995) and detritus is assigned to trophic level one. The food chain or The strength or importance of a trophic relation- ‘Lindeman spine’ that results from these proce- ship cannot be assumed equivalent for all web dures (Fig. 14.5b) is open to comparison with members. A consumer will be a strong interactor ‘spines’ from other systems and many marine and if, in its absence, pronounced changes ensue. Re- freshwater assemblages have been compared as a moval of a weakly interacting species will yield no result (Pauly and Christensen, Chapter 10, Volume or slight change (Paine 1980). 2). Confidence in linear representations of food Although the intent is often to identify large- chains has been raised by strong positive correla- scale characteristics that are of interest to environ- tions found between d15N and body size across mental managers, most food-web data are derived whole trawl assemblages in the North Sea from measurements and manipulations at spatial (Jennings et al. 2001, 2002). The relationship sug- scales that are extremely small (Menge 1992, 1997; gests that body size not taxonomy primarily drives Menge et al. 1994; Wootton 1994). The large-scale predator–prey relations in the North Sea, but this implications of the small-scale observations re- needs to be tested at other locations. main essentially untested. Major indirect effects If food webs constitute useful maps of ecologi- of perturbations characterized from small-scale cal communities (Paine 1988), a major question studies, such as predator controls on prey abun- concerns what they show and where they lead. The dance, trophic cascades and keystone predation, fact is that the existence of a feeding relationship seem to be little known from empirical work at between two organisms does not necessarily large spatial scales (Estes and Duggins 1995; indicate that they interact strongly. Changes in Jennings and Kaiser 1998), whether in benthic one compartment need not lead to accompanying systems (Pinnegar et al. 2000) or pelagic systems changes in the other (Paine 1980, 1992). A major (Verity and Smetacek 1996). reason for this is that predation is only one source Experimentation is the standard scientific ap- of mortality (Kaiser and Jennings, Chapter 16, proach to establishing the strength of connections Volume 2). All biological populations fluctuate in between components in nature, such as the selec- abundance, and the degree of fluctuation may de- tive exclusion experiments of Paine (1992), but at pend on an array of density-dependent and density- large scales in the marine environment rigorous independent factors such as recruitment success, experimentation remains more or less impossible predation, pathogens and the consequences of because of ethical, logistical and financial con- environmental disturbance. An implication of this straints. An alternative is to study the conse- basic fact, combined with the reticulate nature quences of large-scale perturbations arising from of food webs, is that a signal initiated at a point management decisions (Walters’ 1986 concept of Marine Food Webs 313

(a)Exploitation (b) Killer whales p p

Sheephead Spiny Sea otters Seals Sea Starfish Fig. 14.6 lobster sealions otters Trophic cascades p inferred for kelp beds in (a) p p p p California and (b) Alaska. Arrows indicate feeding on prey Sea urchins Sea urchins organisms; h, herbivory; p, predation. (Source: from Pinnegar h h et al. 2000; for primary sources see text.) Kelp Kelp adaptive management) or developments such as macroalgal assemblages (Estes et al. 1978), al- those of fisheries or marine protected areas, or to though a starfish (Pycnopodia helianthoides) also look at responses following climatic events. Such plays a role in determining this structure (Duggins large-scale work has identified a number of cases of 1983). Where sea otters have increased in abun- interactions involving three or more trophic lev- dance, sea urchins have declined, in contrast to els, nearly all of which are based on hard substrata, sites with unchanged sea otter populations where namely rock and coral reef (Pinnegar et al. 2000). sea urchin numbers have remained relatively Three case studies will be used here to characterize steady (Estes and Duggins 1995). In large areas of the dominant types of interaction that have been western Alaska, however, a recent decline in sea found in benthic systems. otter abundance has been attributed to predation Population explosions of the sea urchins by killer whales, predation which had not been Strongylocentrotus franciscanus and S. purpura- observed before 1991 (Estes et al. 1998) (Fig. 14.6b). tus have been known for 30 years to occur in the Since the sea otter is a major focus of conservation kelp forest of the California coast (North and work, an implication is that, at least where it is not Pearse 1969). The spiny lobster Panulirus inter- vulnerable to killer whales, protection of this ani- ruptus (Tegner and Dayton 1981; Tegner and Levin mal will have consequences for the entire system, 1983), sea otter Enhydra lutris (Lowry and Pearse although the outcome is likely to be influenced 1973) and sheephead wrasse Semicossyphus pul- spatially and temporally by the intensity of sea cher (Cowen 1983) have all been implicated as urchin recruitment (Reynolds et al., Chapter 15, biological agents controlling the sea urchins (Fig. Volume 2). 14.6a). Changes in abundance of the sea urchins In the Caribbean, many coral reefs became over- in turn seem to affect abundance of macroalgae grown with macroalgae in the early 1980s and have (Cowen et al. 1982). The evidence is that there are remained so ever since (Hughes 1994). Evidence strong interactions among large predators, major from large-scale correlations in the Caribbean and herbivores and the structure of the macroalgal also comparisons between protected and unpro- community. An implication of this is that ex- tected areas in Kenya suggest that intensive fish- ploitation of the otter, lobster or wrasse will have ing has led to widespread increase in sea urchin important consequences for the entire system densities (Hay 1984; McClanahan 1994) through (Juanes et al., Chapter 12, this volume; Kaiser and the removal of urchin predators and also of com- Jennings, Chapter 16, Volume 2). peting herbivorous fishes (McClanahan and Shafir In the Aleutian Islands of Alaska, islands with 1990) (Fig. 14.7). On shallow Caribbean reefs sea otters have fewer sea urchins (Strongylocentro- where sea urchins were abundant, a pathogen led tus polyacanthus) than those without (Estes and to large-scale decimation of the urchins in 1982–3 Palmisano 1974). The latter tend to have abundant and was the immediate cause of the macroalgal 314 Chapter 14

Exploitation 14.4 IMPLICATIONS OF Large FOOD WEBS AND carnivorous TROPHODYNAMICS fishes p FOR FISH AND p FISHERIES SCIENCE

Sea urchins Grazing There has been debate amongst ecologists over c fishes whether bottom-up forces, such as primary pro- h h ductivity and nutrient availability, or top-down forces, such as predators and consequently fishing, Algae Corals c predominate in communities, and thus whether ‘little things’ (Wilson 1987) or ‘big things’ Fig. 14.7 Relationships among predators, grazers, (Terborgh 1988) exert the most control over macroalgae and corals inferred for reefs on the north coast of Jamaica and in Kenya. Arrows indicate feeding natural systems (Hunter and Price 1992; Persson, on prey organisms; h, herbivory; p, predation; Chapter 15, this volume; Pauly and Christensen, c, competition. (Source: from Pinnegar et al. 2000; for Chapter 10, Volume 2). primary sources see text.) The generally weak evidence for far-reaching effects such as trophic cascades has often been taken as a sign that most marine food webs are overgrowth (Lessios 1988; Hughes 1994), but in driven ‘bottom-up’ (Strong 1992) by carbon fluxes deeper areas correlative and experimental evi- from primary producers (Ryther 1969), and this dence points to grazing fishes being important con- was partly because grazers appeared not to be lim- trollers of macroalgae (Morrison 1988; Williams ited by food supply in the pelagic zone and on reefs. and Polunin 2001). The role of fish may be moder- Greater understanding, at least in the pelagic zone, ated by spatial escapes from herbivory following has indicated that grazing is in fact more intense loss of coral cover through diseases, bleaching than previously thought (Verity and Smetacek and cyclone impacts (Fig. 14.7). Thus, even where 1996), while on reefs food limitation of grazers can- strong evidence for predation controls is deduced, not be excluded because much of the primary pro- other factors that are not explicit in food webs can duction may be accounted for by export (Polunin prove important determinants of community and Klumpp 1992). Consequently, faced with such structure and are likely to influence any future uncertainty, Hunter and Price (1992) offered the change of state in the system. sensible advice that ecologists not ask ‘Do re- Leontief (1951) developed a method to assess sources or predators regulate this particular popu- the importance of direct and indirect interactions lation?’ but rather should try to assess what factors in the economy of the USA, using what has since might moderate resource limitation and predation been called the Leontief matrix. This approach in the system, determining when and where each was introduced to ecology by Hannon (1973) and of these factors will predominate (Kaiser and Hannon and Joiris (1989). Using this method it be- Jennings, Chapter 16, Volume 2 discuss this issue comes possible to assess the effect that changes in with particular regard to the effects of fishing). the biomass of a particular group will have on the Ecological systems are extraordinarily complex biomass of other groups in the system. Ulanowicz and several models have been proposed that link and Puccia (1990) developed this approach and the strength and dynamics of trophic interactions it is now widely utilized as a routine within with the overall stability of communities (re- the Ecopath modelling system (Pauly and viewed in Pimm 1991). Data on natural food webs Christensen, Chapter 10, Volume 2). have generally indicated that these are character- Marine Food Webs 315 ized by many weak interactions and only a few ing and, conversely, the protection of target species strong ones (Paine 1992). May (1973) showed that in marine protected areas (MPAs) (Polunin, Chapter ‘complex’ ecological systems tend to be unstable. 14, Volume 2). For example, it can be predicted that Since complexity will eventually cause a system in MPAs the following conditions will prevail. to fall apart, with strong fluctuations in abundance 1 Trophic level of predators normally targeted and eventual loss of species, the view that held by intensive fishing will tend to increase because sway in the ecological literature for nearly two their average size will tend to be greater and thus decades was that natural systems tend to be quite they will be able to feed on larger prey (Juanes et al., simple and dominated by strong interactions. Chapter 12, this volume), and because large prey However, McCann et al. (1998) have demon- organisms, also targeted by fishing, will tend to be strated that the multitude of weak links that exist more abundant; thus fish of the same size will be in natural systems are important in promoting able to feed more often at higher trophic levels. community persistence and stability by damping 2 The range of trophic levels at which such large oscillations caused by strong interactions between predators will be able to feed, a form of omnivory, consumers and resources and thus preventing the will tend to be greater (Diehl 1992) because the loss of species (Polis 1998). Consequently, strong larger an animal, the greater the range of sizes of interactions may not in fact be as dominant in prey it can potentially eat. natural systems as many originally assumed they 3 Ontogenetic dietary changes, which many were, since food webs dominated by strong inter- fishes exhibit, will occur later if they are size- actions may be more prone to instability (McCann related because there will be more competition for et al. 1998), and may thus be less persistent, than resources, leading to slow growth. The same size those containing many weak linkages. It thus be- will be reached at greater mean age (Persson comes imperative to understand better the extent and Hansson 1999; Mittelbach, Chapter 11, this of strong and weak links in systems. volume). The reefs and other rocky habitats for which As yet, these ideas remain untested. However, strong interactions seem most convincingly estab- it is generally assumed in many studies and indeed lished are scarcely amenable to anything other than in most existing fisheries models that prey are con- small-scale artisanal fisheries. On the other hand, sumed in direct proportion to their abundance in the effects of industrial-scale fisheries are likely to the environment and in relation to the fraction be greater, yet understanding of such effects is in that they represent of the total available prey bio- many respects more primitive (Kaiser and Jennings, mass. This is known in foraging theory as a type II Chapter 16, Volume 2). This seems to be because functional response (Holling 1959; Juanes et al., the large-scale fisheries have scarcely yet been Chapter 12, this volume). A type II functional re- amenable to the spatial comparisons that have sponse implies that consumption rate rises in yielded valuable data on the effects of small-scale proportion to prey density but then gradually fisheries (Jennings and Kaiser 1998). However, decelerates until a plateau is reached. However, responses such as an increase in the biomass of considerable temporal variation has been revealed cephalopods and particularly squid in intensively in the diets of many predatory fish species fished soft-bottom demersal systems do seem to be (Greenstreet et al. 1998; Righton et al. 2001) and a worldwide phenomenon in response to fishing little is known of the actual relationships that (Caddy 1983; Gulland and Garcia 1984; Pauly 1985, exist between the diets of fish and the abundance 1988; Pipitone et al. 2000). of their prey at different temporal and spatial scales Even though the evidence for top-down controls (Greenstreet et al. 1998). For example, predators is not as strong and predictable as might be ex- might switch between different prey species when pected, the understanding of feeding and food webs a particular prey is at low density. The upshot of can nevertheless be used as the basis for making such behaviour (termed a type III response) would several predictions as to the systemic effects of fish- be a sigmoidal relationship between the number of 316 Chapter 14 prey consumed and prey density (Juanes et al., likely to indicate the impacts of particular preda- Chapter 12, this volume). The implication tors. Thus, food webs map ecological communities would be that many existing fisheries models, (Paine 1988) in only a very restricted sense; such as multispecies virtual population analysis nonetheless they are maps of complex systems (Shepherd and Pope, Chapter 7, Volume 2), greatly about which there would otherwise be very little overestimate predation on species that are rare. basis for understanding or prediction.

14.5 CONCLUSIONS ACKNOWLEDGEMENTS

The science involved is far from perfect, but the de- We are grateful to the UK Department for Interna- mands for scientifically based management of the tional Development, Natural Environment Re- coastal zone are growing rapidly and have already search Council and Foundation for Environmental long since surpassed the supply of information Conservation (NVCP) and Fisheries Society of the supported by conventional science. Many investi- British Isles (JKP) for funding, and Villy Chris- gators feel themselves caught between the needs tensen, Dave Raffaelli, Nick Dulvy and Trevor of the world and the discipline of the science. On Willis for useful comments on this chapter. the one hand, the principles of interactions among species within communities are beginning to be understood better and the methods have improved REFERENCES for elucidating them further. On the other hand, it is an understanding of what these principles mean Aebischer, N.J., Coulson, J.C. and Colebrook, J.M. (1990) in the functioning of systems at large scales that is Parallel trends across four marine trophic levels and weather. Nature 347, 753–5. noticeably lacking. For example, the word ‘tropho- Alverson, D.L., Longhurst, A.R. and Gulland, J.A. (1970) dynamics’ used as a label for this field of scientific How much food from the sea? Science 168, 503–5. activity is not a good indicator; a relatively large Azam, F., Fenchel, T., Field, J.G., Meyer-Reil, L.A. and amount is known about the trophic links but Thingstad, F. (1983) The ecological role of water- almost nothing is known of the dynamics. The column microbes in the sea. Marine Ecology Progress data on which food webs are based indicate fluxes, Series 10, 257–63. yet these fluxes are very poorly quantified and Botsford, L.W., Castilla, J.C. and Peterson, C.H. (1997) The management of fisheries and marine ecosystems. aggregation of species-level data into trophic Science 277, 509–15. groups subsumes much detail, a great deal of Briand, F. and Cohen, J.E. (1987) Environmental corre- which is important for the understanding of lates of food chain length. Science 238, 956–60. the functioning of whole systems (Pauly and Caddy, J.F. (1983) The cephalopods: factors relevant to Christensen, Chapter 10, Volume 2). Magnitudes their population dynamics and to the management of of material fluxes are rarely known; where they stocks. FAO Fisheries Technical Paper 231, 416–52. Christensen, N.L., Bartuska, A.M., Brown, J.H. et al. are, they are not good predictors of the relative (1996) The report of the Ecological Society of America strength of interactions when comparing different Committee on the scientific basis for ecosystem man- trophic levels. For example, the fluxes at the bases agement. Ecological Applications 6, 665–91. of food webs that are large, say between primary Cohen, J.E. and Newman, C.M. (1988) Dynamic basis of producers and grazers, in absolute terms do not food web organization. Ecology 69, 1655–64. indicate stronger interactions than between top Cortés, E. (1999) Standardized diet compositions and predators and their prey. Rather, the fluxes are trophic levels of sharks. ICES Journal of Marine Sci- ence 56, 707–17. more predictive of interaction strength within Cowen, R.K. (1983) The effect of sheephead (Semicossy- trophic levels and are especially predictive within phus pulcher) predation on sea urchin (Strongylocen- trophic groups, where different fluxes, such as trotus fransiscanus) populations: an experimental consumption rates by different predators, are most analysis. Oecologia 58, 249–55. Marine Food Webs 317

Cowen, R.K., Agegian, C.R. and Foster, M.S. (1982) The Goldwasser, L. and Roughgarden, J. (1997) Sampling maintenance of community structure in a central effects and the estimation of food-web properties. California giant kelp forest. Journal of Experimental Ecology 78, 41–54. Marine Biology and Ecology 64, 189–201. Greenstreet, S.P.R., McMillan, J.A. and Armstrong, E. Cushing, D.H. (1975) Marine Ecology and Fisheries. (1998) Seasonal variation in the importance of pelagic Cambridge: Cambridge University Press. fish in the diet of piscivorous fish in the Moray Firth, Daan, N. (1989) Database report of the stomach sampling NE Scotland: a response to variation in prey abun- project 1981. ICES Cooperative Research Report dance? ICES Journal of Marine Science 55, 121–33. 164. Copenhagen: International Council for the Explo- Gulland, J.A. (1970) The Fish Resources of the Ocean. ration of the Sea. FAO Fisheries Technical Paper 97. Rome: FAO. Deb, D. (1997) Trophic uncertainty vs parsimony in food Gulland, J.A. and Garcìa, S. (1984) Observed patterns in web research. Oikos 78, 191–4. multispecies fisheries. In: R.M. May (ed.) Exploitation DeNiro, M.J. and Epstein, S. (1978) Influence of diet of Marine Communities. Berlin: Springer-Verlag, pp. on the distribution of carbon isotopes in animals. 155–90. Geochimica et Cosmochimica Acta 42, 495–506. Haeuber, R. and Franklin, J. (1996) Perspectives on Diehl, S. (1992) Fish predation and benthic community ecosystem management. Ecological Applications 6, structure: the role of omnivory and habitat com- 692–3. plexity. Ecology 73, 1646–61. Hagen, W., Katter, G. and Graeve, M. (1995) On the lipid Doucett, R.R., Power, G., Barton, D.R., Drimmie, R.J. biochemistry of polar copepods: compositional differ- and Cunjak, R.A. (1996) Stable isotope analysis of nu- ences in the Antarctic calanoids Euchaeta antarctica trient pathways leading to Atlantic salmon. Canadian and Euchirella rostromagma. Marine Biology 123, Journal of Fisheries and Aquatic Sciences 53, 2058–66. 451–7. Duggins, D.O. (1983) Starfish predation and the creation Hall, S.J. and Raffaelli, D.G. (1993) Food webs: theory and of mosaic patterns in a kelp-dominated community. reality. Advances in Ecological Research 24, 187–239. Ecology 64, 1610–19. Hannon, B. (1973) The structure of ecosystems. Journal Elton, C.S. (1927) Animal Ecology. London: Sidgwick & of Theoretical Biology 41, 535–46. Jackson. Hannon, B. and Joiris, C. (1989) A seasonal analysis of the Estes, J.A. and Duggins, D.O. (1995) Sea otters and kelp southern North Sea ecosystem. Ecology 70, 1916–34. forests in Alaska: generality and variation in a commu- Hardy, A. (1959) The Open Sea: its Natural History. Part nity ecological paradigm. Ecological Monographs 65, II: Fish and Fisheries. London: Collins. 75–100. Harrigan, P., Zieman, J.C. and Macko, S.A. (1989) The Estes, J.A. and Palmisano, J.F. (1974) Sea otters: their role base of nutritional support for the Gray Snapper in structuring nearshore communities. Science 185, (Lutjanus griseus): an evaluation based on a combined 1058–60. stomach content and stable isotope analysis. Bulletin Estes, J.A., Smith, N.A. and Palmisano, J.F. (1978) Sea of Marine Science 44, 65–77. otter predation and community organization in the Hay, M.E. (1984) Patterns of fish and urchin grazing on western Aleutian Islands, Alaska. Ecology 59, 822–33. Caribbean coral reefs: are previous results typical? Estes, J.A., Tinker, M.T., Williams, T.M. and Doak, D.F. Ecology 65, 446–54. (1998) Killer whale predation on sea otters linking Hentschel, B.T. (1998) Intraspecific variations in d13C oceanic and nearshore ecosystems. Science 282, 473–6. indicate ontogenetic diet changes in deposit-feeding FAO (1995) The State of the World Fisheries and Aqua- polychaetes. Ecology 79, 1357–70. culture. Rome: FAO. Hesslein, R.H., Capel, M.J., Fox, D.E. and Hallard, K.A. Fraser, A.J., Sargent, J.R., Gamble, J.C. and Seaton, D.D. (1991) Stable isotopes of sulfur, carbon, and nitrogen as (1989) Formation and transfer of fatty acids in an en- indicators of trophic level and fish migration in the closed marine food chain comprising phytoplankton, Lower Mackenzie River Basin, Canada. Canadian zooplankton and herring (Clupea harengus L.) larvae. Journal of Fisheries and Aquatic Sciences 48, 2258–65. Marine Chemistry 27, 1–18. Hesslein, R.H., Hallard, K.A. and Ramal, P. (1993) Re- Frid, C.L.J. and Hall, S.J. (1999) Inferring changes in placement of sulphur, carbon and nitrogen of growing North Sea benthos from fish stomach analysis. Marine broad whitefish (Coregonus nasus) in response to a Ecology Progress Series 184, 183–8. change in diet traced by d13C, d15N and d34S. Canadian Gannes, L.Z., Obrien, D.M. and delRio, C.M. (1997) Journal of Fisheries and Aquatic Sciences 50, 2071–6. Stable isotopes in animal ecology: assumptions, Hobson, K.A. (1999) Tracing origins and migration of caveats, and a call for more laboratory experiments. wildlife using stable isotopes: a review. Oecologia 120, Ecology 78, 1271–6. 314–26. 318 Chapter 14

Holling, C.S. (1959) Some characteristics of simple types signature of whole Atlantic cod (Gadus morhua). of predation and parasitism. Canadian Entomologist Canadian Journal of Fisheries and Aquatic Sciences 91, 385–98. 55, 1378–86. Hughes, T.P. (1994) Catastrophes, phase shifts, and Kline, T.C. and Pauly, D. (1998) Cross-validation of large-scale degradation of a Caribbean coral reef. trophic level estimates from a mass-balance model of Science 265, 1547–51. Prince William Sound using 15N/14N data. In: F. Funk, Hunter, M.D. and Price, P.W. (1992) Playing chutes and T.J. Quinn II, J. Heifetz, J.E. Powers, J.F. Schweigert, P.J. ladders: heterogeneity and the relative roles of bottom- Sullivan and C.-I. Zhang (eds) Fishery Stock Assess- up and top-down forces in natural communities. ment Models. Fairbanks, AK: University of Alaska, pp. Ecology 73, 724–32. 693–702. Hyslop, E.J. (1980) Stomach contents analysis: a review of Lasker, R. (1988) Food chains and fisheries: an assess- methods and their application. Journal of Fish Biology ment after 20 years. In: B.J. Rothschild (ed.) Toward 17, 411–29. a Theory on Biological–Physical Interactions in the Iverson, S.J. (1993) Milk secretion in marine mammals in World Ocean. Dordrecht: Kluwer Academic, pp. relation to foraging: can milk fatty acids predict diet? 173–82. Symposia of the Zoological Society of London 66, Leontief, W.W. (1951) The Structure of the US Economy. 263–91. New York: Oxford University Press. Iverson, S.J., Oftedal, O.T., Bowen, W.D., Boness, D.J. and Lessios, H.A. (1988) Mass mortality of Diadema antil- Sampugna, J. (1995) Prenatal and postantal transfer larum in the Caribbean: what have we learned? An- of fatty acids from mother to pup in the hooded seal. nual Review of Ecology and Systematics 19, 371–93. Journal of Comparative Physiology B 165, 1–12. Lindeman, R.L. (1942) The trophic-dynamic aspect of Iverson, S.J., Frost, K.J. and Lowry, L.F. (1997) Fatty acid ecology. Ecology 23, 399–418. signatures reveal fine scale structure of foraging distri- Lindsay, D.J., Minagawa, M., Mitani, I. and Kawaguchi, bution of harbor seals and their prey in Prince William K. (1998) Trophic shift in the Japanese anchovy Sound, Alaska. Marine Ecology Progress Series 151, Engraulis japonicus in its early life history stages as 255–71. detected by stable isotope ratios in Sagami Bay, Central Jennings, S. and Kaiser, M.J. (1998) The effects of fishing Japan. Fisheries Science 64, 403–10. on marine ecosystems. Advances in Marine Biology Logan, M.S., Iverson, S.J., Ruzzante, D.E., Walde, S.J., 34, 201–352. Macchi, P.J., Alonso, M.F. and Cussac, V.E. (2000) Jennings, S., Reñones, O., Morales-Nin, B., Polunin, Long term diet differences between morphs in N.V.C., Moranta, J. and Coll, J. (1997) Spatial variation trophically polymorphic Percichthys trucha (Pisces: in the 15N and 13C stable isotope composition of plants, Percichthyidae) populations from the southern invertebrates and fishes on Mediterranean reefs: impli- Andes. Biological Journal of the Linnean Society 69, cations for the study of trophic pathways. Marine Ecol- 599–616. ogy Progress Series 146, 109–16. Longhurst, A. (1998) Ecological Geography of the Sea. Jennings, S., Pinnegar, J.K., Polunin, N.V.C. and Boon, T. San Diego, CA: Academic Press. (2001) Weak cross-species relationships between body Longhurst, A., Sathyendranath, S., Platt, T. and size and trophic level belie powerful size-based trophic Caverhill, C. (1995) An estimate of global primary structuring in fish communities. Journal of Animal production in the ocean from satellite radiometer data. Ecology 70, 934–44. Journal of Plankton Research 17, 1245–71. Jennings, S., Pinnegar, J.K., Polunin, N.V.C. and Warre, Lowry, L.F. and Pearse, J.S. (1973) Abalones and sea K.J. (2002) Linking size-based and trophic analyses urchins in an area inhabited by sea otters. Marine Biol- of benthic community structure. Marine Ecology ogy 23, 213–19. Progress Series 226, 77–85. Magnusson, K.G. (1995) An overview of the multispecies Jones, G.P. (1991) Postrecruitment processes in the eco- VPA: theory and applications. Reviews in Fish Biology logy of coral reef fish populations: a multifactorial and Fisheries 5, 195–212. perspective. In: P.F. Sale (ed.) The Ecology of Fishes May, R.M. (1973) Stability and Complexity in Model on Coral Reefs. San Diego, CA: Academic Press, pp. Ecosystems. Princeton, NJ: Princeton University 294–328. Press. Keegan, W.F. and DeNiro, M.J. (1988) Stable carbon- and McCann, K., Hastings, A. and Huxel, G.R. (1998) Weak nitrogen-isotope ratios of bone collagen used to study trophic interactions and the balance of nature. Nature coral-reef and terrestrial components of prehistoric 395, 794–8. Bahamian diet. American Antiquity 53, 320–36. McClanahan, T.R. (1994) Kenyan coral reef lagoon fish: Kirsch, P.E., Iverson, S.J., Bowen, W.D., Kerr, S.R. and effects of fishing, substrate complexity and sea Ackman, R.G. (1998) Dietary effects on the fatty acid urchins. Coral Reefs 13, 231–41. Marine Food Webs 319

McClanahan, T.R. and Shafir, S.H. (1990) Causes and Pauly, D., Trites, A.W., Capuli, E. and Christensen, V. consequences of sea urchin abundance and diversity in (1998b) Diet composition and trophic levels of marine Kenyan coral reef lagoons. Oecologia 83, 362–70. mammals. ICES Journal of Marine Science 55, 467–81. Menge, B.A. (1992) Community regulation: under what Pauly, D., Christensen, V. and Walters, C. (2000) Ecopath, conditions are bottom-up factors important on rocky Ecosim and Ecospace as tools for evaluating ecosystem shores? Ecology 73, 755–65. impact of fisheries. ICES Journal of Marine Science 57, Menge, B.A. (1997) Detection of direct versus indirect ef- 697–706. fects: were experiments long enough? American Natu- Pedersen, L., Jensen, H.M., Burmeister, A. and Hansen, ralist 149, 801–23. B.W. (1999) The significance of food web structure for Menge, B.A., Berlow, E.L., Blanchette, C.A., Navarette, the condition and tracer lipid content of juvenile snail S.A. and Yamada, S.B. (1994) The keystone concept: fish (Pisces: Liparis spp) along 65–72 degrees N off variation in interaction strength in a rocky intertidal West Greenland. Journal of Plankton Research 21, habitat. Ecological Monographs 64, 249–86. 1593–611. Minagawa, M. and Wada, E. (1984) Stepwise enrichment Persson, A. and Hansson, L.A. (1999) Diet shift in fish fol- of 15N along food chains: further evidence and the lowing competitive release. Canadian Journal of Fish- relation between d15N and animal age. Geochimica et eries and Aquatic Sciences 56, 70–8. Cosmochimica Acta 48, 1135–40. Pimm, S.L. (1982) Food Webs. London: Chapman & Hall. Monteiro, P.M.S., James, A.G., Sholto-Douglas, A.D. and Pimm, S.L. (1988) Energy flow and trophic structure. In: Field, J.G. (1991) The d13C trophic position isotope L.R. Pomeroy and J.J. Alberts (eds) Concepts of Ecosys- spectrum as a tool to define and quantify carbon path- tem Ecology. Berlin: Springer-Verlag, pp. 263–78. ways in marine food webs. Marine Ecology Progress Se- Pimm, S.L. (1991) The Balance of Nature: Ecological ries 78, 33–40. Issues in the Conservation of Species and Communi- Morowitz, H.J. (1991) The Thermodynamics of Pizza: ties. Chicago: University of Chicago Press. Essays on Science and Everyday Life. New Brunswick, Pimm, S.L., Lawton, J.H. and Cohen, J.E. (1991) Food web NJ: Rutgers University Press. patterns and their consequences. Nature 350, 669–74. Morrison, D. (1988) Comparing fish and urchin grazing in Pinnegar, J.K. and Polunin, N.V.C. (1999) Differential shallow and deeper coral reef algal communities. Ecol- fractionation of d13C and d15N among fish tissues: ogy 69, 1367–82. implications for the study of trophic interactions. North, W.J. and Pearse, J.S. (1969) Sea urchin population Functional Ecology 13, 225–31. explosion in southern California coastal waters. Sci- Pinnegar, J.K. and Polunin, N.V.C. (2000) Contributions ence 167, 209. of stable-isotope data to elucidating food webs of Odum, E.P. (1968) Energy flow in ecosystems: a historical Mediterraenan rocky littoral fishes. Oecologia 122, review. American Zoologist 8, 11–18. 399–409. Paine, R.T. (1980) Food webs: linkage, interaction Pinnegar, J.K., Polunin, N.V.C., Francour, P. et al. (2000) strength and community infrastructure. Journal of Trophic cascades in benthic marine ecosystems: Animal Ecology 49, 667–85. lessons for fisheries and protected-area management. Paine, R.T. (1988) Food webs: road maps of interactions Environmental Conservation 27, 179–200. or grist for theoretical development? Ecology 69, Pipitone C., Badalamenti, F., D’Anna, G. and Patti, B. 1648–54. (2000) Fish biomass increase after a four-year trawl ban Paine, R.T. (1992) Food-web analysis through field mea- in the Gulf of Castellammare (NW Sicily, Mediter- surement of per capita interaction strength. Nature ranean Sea). Fisheries Research 48, 23–30. 355, 73–5. Polis, G.A. (1998) Stability is woven by complex webs. Pauly, D. (1985) The population dynamics of short-lived Nature 395, 744–5. species, with emphasis on squids. Northwest Atlantic Polis, G.A. and Strong, D.R. (1996) Food web complexity Fisheries Organization Scientific Council Studies 9, and community dynamics. American Naturalist 147, 143–54. 813–46. Pauly, D. (1988) Fisheries research and the demersal fish- Polovina, J.J., Mitchum, G.T., Graham, N.E., Craig, M.P., eries of Southeast Asia. In: J.A. Gulland (ed.) Fish Popu- DeMartini, E.E. and Flint, E.E. (1994) Physical and bio- lation Dynamics, 2nd edn. Chichester: John Wiley & logical consequences of a climatic event in the Central Sons, pp. 329–48. Pacific. Fisheries Oceanography 3, 15–21. Pauly, D. and Christensen, V. (1995) Primary production Polunin, N.V.C. and Klumpp, D.W. (1992) Algal supply required to sustain global fisheries. Nature 374, 255–7. and grazer demand in a very productive coral-reef zone. Pauly, D., Christensen, V., Dalsgaard, J., Froese, R. and Journal of Experimental Marine Biology and Ecology Torres, F. (1998a) Fishing down marine food webs. Sci- 164, 1–15. ence 279, 860–3. Pond, D.W., Dixon, D.R., Bell, M.V., Fallick, A.E. and 320 Chapter 14

Sargent, J.R. (1997) Occurrence of 16:2(n - 4) and 18: Diets and long-term changes in d15N and d13C values in 2(n - 4) fatty acids in the lipids of the hydrothermal Northern Fulmars Fulmarus glacialis from two north- vent shrimps Rimicaris exoculata and Alvinocaris east Atlantic colonies. Marine Ecology Progress Series markensis: nutritional and trophic implications. Ma- 125, 3–11. rine Ecology Progress Series 156, 167–74. Ulanowicz, R.E. (1986) Growth and Development: Raclot, T., Groscolas, R. and Cherel, Y. (1998) Fatty acid Ecosystems Phenomenology. New York: Springer- evidence for the importance of myctophid fishes in the Verlag. diet of king penguins, Aptenodytes patagonicus. Ma- Ulanowicz, R.E. (1995) Ecosystem trophic foundations: rine Biology 132, 523–33. Lindeman exonerata. In: B.C. Patten and S.E. Jorgensen Righton, D., Metcalfe, J. and Connolly, P. (2001) Different (eds) Complex Ecology: the Part–Whole Relation in behaviour of North and Irish sea cod. Nature 411, 156. Ecosystems. Englewood Cliffs, NJ: Prentice-Hall, pp. Robertson, D.R. (1982) Fish feces as fish food on a Pacific 549–60. coral reef. Marine Ecology Progress Series 7, 253–65. Ulanowicz, R.E. and Puccia, C.J. (1990) Mixed trophic Ryther, J.H. (1969) Photosynthesis and fish production in impacts in ecosystems. Coenoses 5, 7–16. the sea. Science 166, 72–6. van der Veer, H.W., Feller, R.J., Weber, A. and Witte, J.I.J. Sargent, J.R. (1976) The structure, function and metabo- (1998) Importance of predation by crustaceans upon bi- lism of lipids in marine organisms. In: D.C. Malins and valve spat in the intertidal zone of the Dutch Wadden J.R. Sargent (eds) Biochemical and Biophysical Per- Sea as revealed by immunological assays of gut con- spectives in Marine Biology. London: Academic Press, tents. Journal of Experimental Marine Biology and pp. 150–212. Ecology 231, 139–57. Sargent, J.R. (1978) Marine wax esters. Science Progress Vander Zanden, M.J., Cabana, G. and Rasmussen, J.B. Oxford 65, 437–58. (1997) Comparing the trophic position of littoral fish Sargent, J.R., Parkes, R.J., Muller-Harvey, I. and estimated using stable nitrogen isotopes (15N) and di- Henderson, R.J. (1988) Lipid biomarkers in marine etary data. Canadian Journal of Fisheries and Aquatic ecology. In: M.A. Sleigh (ed.) Microbes in the Sea. Sciences 54, 1142–58. Chichester: Ellis Horwood, pp. 119–38. Verity, P.G. and Smetacek, V. (1996) Organism life cycles, Slobodkin, L.B. (1961) Growth and Regulation of predation, and the structure of marine pelagic ecosys- Aquatic Populations. New York: Holt, Rinehart & tems. Marine Ecology Progress Series 130, 277–93. Winston. Wainright, S.C., Fogarty, M.J., Greenfield, R.C. and Fry, Smith, R.J., Hobson, K.A., Koopman, H.N. and Lavigne, B. (1993) Long-term changes in the Georges Bank food D.M. (1996) Distinguishing between populations of web: trends in stable isotopic compositions of fish fresh- and salt-water harbour seals (Phoca vitulina) scales. Marine Biology 115, 481–93. using stable-isotope ratios and fatty acid profiles. Walters, C.J. (1986) Adaptive Management of Renew- Canadian Journal of Fisheries and Aquatic Sciences able Resources. New York: Macmillan (also reprinted 53, 272–9. by the Fisheries Centre, University of British Colum- Sterner, R.W., Bajpai, A. and Adams, T. (1997) The bia, Vancouver, Canada). enigma of food chain length: absence of theoretical Walters, C.J. and Holling, C.S. (1990) Large-scale man- evidence for dynamic constraints. Ecology 78, 2258– agement experiments and learning by doing. Ecology 62. 71, 2060–8. Strong, D.R. (1992) Are trophic cascades all wet? Differ- Walters, C., Christensen, V. and Pauly, D. (1997) Struc- entiation and donor-control in species ecosystems. turing dynamic models of exploited ecosystems from Ecology 73, 747–54. trophic mass-balance assessments. Reviews in Fish Tegner, M.J. and Dayton, P.K. (1981) Population struc- Biology and Fisheries 7, 139–72. ture, recruitment and mortality of two sea urchins Williams, I.D. and Polunin, N.V.C. (2001) Large-scale as- (Strongylocentrotus fransiscanus and S. purpuratus) in sociations between macroalgal cover and grazer bio- a kelp forest. Marine Ecology Progress Series 5, 255–68. mass on mid-depth reefs in the Caribbean. Coral Reefs Tegner, M.J. and Levin, L.A. (1983) Spiny lobsters and 19, 358–66. sea urchins: analysis of a predator–prey interaction. Wilson, E.O. (1987) The little things that run the world Journal of Experimental Marine Biology and Ecology (the importance and conservation of invertebrates). 73, 125–50. Conservation Biology 1, 344–6. Terborgh, J.W. (1988) The big things that run the world: a Wootton, T.J. (1994) Predicting direct and indirect ef- sequel to E.O Wilson. Conservation Biology 2, 402–3. fects: an integrated approach using experiments and Thompson, D.R., Furness, R.W. and Lewis, S.A. (1995) path analysis. Ecology 75, 151–65. 15 Community Ecology of Freshwater Fishes

LENNART PERSSON

15.1 INTRODUCTION mance of different species in relation to abiotic fac- tors such as light and temperature. This treatment Ecology has been defined as the scientific study of can largely be seen as a treatment of the fundamen- the interactions between organisms and their en- tal ecological niche, sensu Hutchinson, of differ- vironments (Krebs 1972). A useful starting point to ent species. The actual niche of a species will be address questions about interactions among or- constrained by ecological interactions with other ganisms in relation to environmental conditions is species. Fish ecology has played a prominent role to study patterns in the distribution of species or in the development of ecological theory regarding functional groups over space and time. Changes niche dynamics and ecological interactions in over space in abundance relationships between general (Nilsson 1963; Svärdson 1976; Werner and fish species or groups of species have been Hall 1977, 1979; Werner 1986; Persson 1988). The observed along several environmental gradients, presence of size structure also means that interac- including temperature, light and productivity tions among fish populations are generally a result (Colby et al. 1972; Leach et al. 1977; Magnuson of a mixture of size-dependent competitive and et al. 1979; Schlosser 1987; Persson et al. 1991, predator–prey interactions (Werner 1986; Persson 1992). The purpose of this chapter is to discuss 1988; Mittelbach and Osenberg 1993; Persson et al. some of these patterns but also to provide a 1999; Juanes et al., Chapter 12, this volume). framework for a mechanistic understanding of the Studies of distribution patterns across systems patterns observed. Because a fundamental charac- show that size-structured interactions are of major teristic of fish populations, as with populations of importance for large-scale community patterns many other species, is that individuals differ sub- (Persson et al. 1991, 1992; Mittelbach and stantially in size, such a treatment has to consider Osenberg 1993; Olson et al. 1995). Changes in fish the implications of size structure for ecological community structure also have consequences for interactions. other trophic components of aquatic food webs In the following, I first consider studies that have (Carpenter et al. 1987; Persson et al. 1992). I consi- focused on large-scale patterns in the distributions der the effects of fish on lower trophic compo- of fish species, including between-continent com- nents, which include a number of pathways such parisons often using multivariate techniques. This as direct predation on zooplankton, effects of fish part shows that a number of basic environmental consumption on nutrient recycling in pelagic envi- variables, such as temperature, light and produc- ronments, and translocation of nutrients between tivity, have major influences on the distribution of habitats (Schindler et al. 1996; Vanni 1996). species across environmental gradients. At the end In a final section I discuss how size structure in of this section I deal with the ecological perfor- fish populations can be formally analysed using a 322 Chapter 15 modelling approach termed ‘physiologically struc- using multivariate techniques have been carried tured population models’ (De Roos 1997; Persson in North American (particularly Wisconsin), et al. 1998). This modelling approach is based Finnish and Swedish lakes (Tonn et al. 1990, 1995; on a two-state-level concept and explicitly links Magnuson et al. 1998). These studies mainly con- individual-level processes such as individual sidered small, often forest, lakes and have sug- growth to population-level processes. I argue that gested that extinction is more important than this modelling approach may form a productive isolation, which influences immigration, in deter- research agenda for increasing our understanding mining community structure (Tonn et al. 1995; of the implications of size structure for the com- Magnuson et al. 1998). This has been related to the munity dynamics of fish. Methods for obtaining fact that small forest lakes have a number of fea- population parameters from size distributions tures that contribute to frequent (relative to immi- are given by Pitcher (Chapter 9, Volume 2). gration) extinctions, such as low pH, seasonally occurring low oxygen levels and piscivores. The degree of isolation was not unimportant but it 15.2 COMMUNITY played a lesser role in observed patterns in species PATTERNS AND BASIC richness and assembly type (species configura- ECOLOGICAL PERFORMANCE tion). However, differences were found between continents. In Finnish lakes, assembly type was 15.2.1 Large-scale explained by isolation to a larger extent than by community patterns each of four extinction factors, which was not the case in Wisconsin lakes (Tonn et al. 1990). Many studies focusing on fish community struc- Magnuson et al. (1998) suggested that the rea- ture have stressed the insular nature of many lakes son that extinction is more important than isola- and the similarities between lakes and islands tion in explaining species richness and assembly (Rahel 1984). This is so particularly for the theory type in small lakes is because immigration rates of island biogeography, which relates species di- are low, often varying between decades to millen- versity to two major processes: immigration and nia. In contrast, extinctions can occur at substan- extinction rates (MacArthur and Wilson 1967). tially shorter time-scales from months up to a few The focus of island biogeography theory has tradi- decades after invasions. Correspondingly, studies tionally been on species–area relationships using in North American lakes show that immigration the slope of the regression between species diver- rates are much higher in larger lakes than in sity and area to analyse community patterns. How- smaller lakes, partly due to human influences. ever, as noted by Magnuson et al. (1998), the slope Thus, lake area determines the importance of ex- of the species–area regression has a variety of possi- tinction rates vs. immigration rates in explaining ble interpretations with respect to the importance species diversity and assembly type. to species diversity of isolation and extinction. In Wisconsin lakes, three major factors were Area per se is also correlated with many isolation isolated as being of prime importance for extinc- and extinction properties of lakes. For example, tion rates: winter-kills due to low oxygen, low pH there is a negative relationship between lake area and predation by pike (Esox lucius) or largemouth and the proportion of lakes without a connecting bass (Micropterus salmoides) (Magnuson et al. stream in small forest lakes (Magnuson et al. 1998). 1998). The frequency of winter-kills is expected To reduce the problem with interpretations and to depend on environmental factors like lake correlations, multivariate approaches have been depth, productivity, latitude and altitude. The advanced to separate the possible roles of immigra- importance of pH in determining species richness tion and extinction in comparative studies (Tonn has become increasingly clear as a result of the et al. 1990; Magnuson et al. 1998). The most exten- acidification of lakes and its amelioration by sive comparative studies on community patterns liming that aims at restoring fish communities Community Ecology of Freshwater Fishes 323

(e.g. Appelberg 1998). Sensitivity to low pH also 100 differs between species: cyprinids and salmonid fishes are more sensitive than, for example, perch 75 (Perca fluviatilis). Extinctions due to predation have been reported from North American lakes, where predators like northern pike, largemouth 50 bass and yellow perch (Perca flavescens) may eliminate populations of small cyprinids. The Species richness 25 extinction probability for sensitive prey fish popu- lations after their invasion of Wisconsin lakes in- habited by piscivores was on the same range as the 0 0 5 10 15 20 25 30 extinction probability due to winter-kills and pH (Magnuson et al. 1998). In contrast, in European Fluvial distance (km) lakes cyprinids are able to persist more readily Fig. 15.1 Changes in species richness with fluival with piscivorous predators because Eurasian distance from mouth in Rio Tortuguero, Costa Rica. cyprinids can grow to a size where they are no (Source: data from Winemiller and Leslie 1992.) longer susceptible to predation (Tonn et al. 1990). Other differences in processes affecting species richness and assembly type have also been found phological diversification with decreasing lati- between continents and regions. As described tude. This increase in morphological diversifica- above, extinction factors were more important tion in high-diversity tropical faunas was, based on than isolation factors for Wisconsin lakes in com- ecomorphological measures, not associated with parison with Finnish lakes (Magnuson et al. 1998). an increase in species packing. However, whether Furthermore, conductivity and lake depth were the greater species richness associated with greater more important in Finnish lakes, whereas pH was niche diversification was a result of an expansion more important in Wisconsin lakes. Swedish lakes of the total range of food and habitat resources used differed from Finnish and North American lakes or due to a finer subdivision of available resources in that geographical factors were of outstanding could not be determined based on ecomorphologi- importance for separating Swedish lakes, whereas cal measures only. abiotic factors and lake morphometry were more important in Finland and in two North American 15.2.2 Basic ecological performance regions (Tonn et al. 1995). Studies of species distributions of fish in streams The performance of different fish species under show that headwater streams with low aquatic different abiotic environmental conditions can, primary productivity generally have lower species among other things, be related to the fact that richness and fish assemblages, characterized by physiological rates are directly influenced by fast-colonizing fish species subject to large fluctu- factors such as temperature, light, salinity, pH ations in local densities (Schlosser 1987). Studies and oxygen and that different species have differ- of tropical systems suggest that these systems ent optima with respect to these factors (Power show the same pattern in changes in species rich- 1990; Brix, Chapter 4, this volume). In the follow- ness with stream order as temperate systems, sug- ing, I consider temperature and light in more gesting that similar mechanisms are operating in detail with respect to the performance of different structuring fish communities in temperate and species. tropical systems (Winemiller and Leslie 1992) (Fig. For poikilotherms like fish, temperature is a 15.1). Complementing these studies, Winemiller major factor affecting overall ecological perfor- (1992) found in a comparative interregional/inter- mance through its effect on energetics (Brett and continental study an increase in species ecomor- Groves 1979). (Effects of temperature on rates of 324 Chapter 15

(a)2 (b) 1.0 Roach Ruffe Perch Perch ) –1

1 0.5

Fig. 15.2 (a) Maximum capture Capture rate (no. s rates of roach and perch feeding on Chaoborus larvae at 15 and 21°C. (Source: data from Persson 1986.) (b) Maximum capture rates of ruffe 0 0.0 15 21 0.02 10 and perch feeding on Chaoborus larvae at 0.02 and 10lx. (Source: Temperature ( C) Illumination (lux) data from Bergman 1988.)

development are reviewed by Jobling in Chapter 5, that the maximum consumption rate of Eurasian this volume, and effects on metabolism by Brix in perch was higher than that of roach (Rutilus Chapter 4, this volume.) Experimental studies rutilus) at temperatures below 18°C, whereas the have shown differences between species in tem- opposite was the case at temperatures above 18°C perature-dependent consumption and growth (Fig. 15.2a). This difference in temperature- performance. Among salmonid fishes, brown dependent foraging ability relates in turn to the trout Salmo trutta have their maximum consump- fact that perch are distributed further north in tion rate and growth rate at higher temperatures Scandinavian lakes and also penetrate deeper in than do Arctic char Salvelinus alpinus (Elliott stratified lakes. The impact of temperature on the 1975; Jobling 1983). Temperature also influences ecological performance of fish led Magnuson et al. the large-scale distributions of species. For exam- (1979) to advance temperature as a dimension ple, Power (1990) found that temperature was along which fish species with different tempera- important in explaining the geographical distribu- ture optima could separate their niches. tion of salmonid species. Within systems, Baltz Species differences have also been shown in re- et al. (1982) showed that temperature was a major sponse to light conditions (Schultz and Northcote determinant of species abundance in streams. A 1972; Bergman 1988). Schultz and Northcote major route by which temperature affects the eco- (1972) described Dolly Varden char (Salvelinus logical performance of fish species is through its malma) as a more efficient forager in dim light effect on foraging rate. Hammar (1998) suggested conditions compared with cutthroat trout (On- that differences in feeding performance of brown corhynchus clarki). Additionally, differences in trout and Arctic char at different temperatures, foraging abilities have been suggested to be impor- particularly the fact that Arctic char may feed in tant for the ecological interactions between Arctic winter under the ice, were important in explaining char and brown trout, with trout feeding on more the distribution of these two species within and visible surface prey and chironomidae pupae at among lakes. Experimental studies have also night while Arctic char mainly feed on zooplank- demonstrated differences in functional responses ton during the same time period (Hammar 1998). of species as a function of temperature (for details Experimentally, Bergman (1988) showed that ruffe of different functional responses see Juanes et al., (Gymnocephalus cernua) had a higher foraging Chapter 12, this volume). Persson (1986) showed rate than perch under poor light conditions, Community Ecology of Freshwater Fishes 325 whereas the reverse was the case under good light lakes, Svärdson (1976) advanced the hypothesis conditions (Fig. 15.2b). This difference in light- that it is the more planktivorous species, the Arctic dependent foraging ability can, in turn, be related char in the brown trout–Arctic char interaction, to the ecological performance of ruffe and perch that will numerically dominate the system. This in lakes that differ in light conditions because of pioneering work by Swedish fish ecologists was different productivities (Bergman 1988). relatively unnoticed by non-fish ecologists and The ecological performance of different species remained so until the early 1980s (Schoener 1982). as a function of abiotic conditions such as tempera- Similar studies on interactive segregation were also ture and light is essentially a description of the fun- carried out on salmonids in North America damental niche of the species. In interactions with (Schultz and Northcote 1972). other species, or size cohorts within the same The presence of interactive segregation was sub- species, the realized niche of a species or size co- stantiated by studies of changes in resource use. In hort will generally shrink to the part of the funda- temperate systems these changes were related to mental niche where individuals have the highest seasonal variation in resource levels or tempera- fitness. Operationally this can, for example, be ture (Nilssson 1960; Persson 1987a). In tropical measured by the attack rate of a particular species systems, interactive segregation has been shown or size cohort on different prey types. The next by comparing the degree of resource partitioning in section covers the implications of ecological dry versus wet seasons (Zaret and Rand 1971). In- interactions with a focus on the size-dependency terestingly, data suggest that piscivorous and non- of these interactions. piscivorous species may show different responses to the wet season. Non-piscivorous species gener- ally show an increase in food overlap in wet sea- 15.3 COMPETITION sons, interpreted as a response to increased food AND PREDATION AS supply that decreased competition (i.e. Zaret and STRUCTURING FORCES Rand 1971). In contrast, Winemiller and Ponwith (1998) found that the diet similarity between 15.3.1 Impact of piscivorous fish species decreased during the wet competitive interactions season when prey dispersed at lower per-unit- area densities. Comparative studies on fish communities were The early studies on ecological interactions in one of the first to document the impact of competi- fish communities were mainly based on compara- tion on species distributions within and among tive studies and the patterns observed were thus systems (Svärdson 1949; Nilsson 1963, 1978). In open to alternative explanations. The pioneering studies of Arctic char and brown trout, Nilsson experimental work by Werner and Hall (1977, demonstrated that the habitat and food use of the 1979) on sunfishes (Lepomis) was a qualitative step two species differed depending on whether they forward in that the presence of competition could were sympatric or allopatric. In lakes where the two be unambiguously demonstrated and also that species lived in allopatry, both of them utilized the a more predictive approach was developed. In a whole water volume. In contrast, in lakes where series of elegant experiments, Werner and Hall they lived in sympatry, Arctic char restricted their showed that the bluegill sunfish (L. macrochirus) habitat use to the pelagic zone, whereas trout used was a superior forager to green sunfish (L. cyanel- the littoral shore habitat. The plasticity observed in lus) and pumpkinseed sunfish (L. gibbosus) in the habitat use of the two species was a result of interac- open-water habitat. They also showed that the tive segregation, in contrast to selective segrega- green sunfish was the superior forager in vegeta- tion where the habitat uses of competing species tion. In a pond experiment where resources in the have become genetically fixed (Nilsson 1978). In a vegetation were depleted over time, Werner and review of species interactions in Scandinavian Hall (1979) predicted that the bluegill sunfish 326 Chapter 15 should be the first to shift to the pelagic habitat and competitor to perch (Persson 1988), young-of-the- the green sunfish the last, a prediction cor- year (YOY) perch were less affected by competition roborated by results (Werner and Hall 1979). The than 1-year-old roach due to the former’s lower studies of the Centrarchidae subsequently evolved metabolic demands (Byström and Garcià-Berthóu towards a more quantitative approach by utilizing 1999). This dependency of the outcome of com- optimal foraging theory to predict net energy petition between different-sized individuals on intake from different habitats and thereby the both size-dependent foraging (which affects attack optimal habitat to be in (Mittelbach, Chapter 11, rate and handling) and metabolic rates illustrates this volume). In a pond experiment lacking how size-structured interactions also lead to predators, predictions of optimal habitat use were asymmetric competitive interactions between largely supported by empirical data (Werner et al. different-sized individuals (Werner 1986; Persson 1983a). In lakes with predators, predictions from et al. 1998). foraging theory were upheld for size classes of The intensity of resource limitation among size bluegills that were not susceptible to predators cohorts of fish may vary with size, causing a (Mittelbach 1981). However, small bluegills that certain stage or size class to form a bottleneck for were susceptible to predation from predators like recruitment whereas other stages/sizes may not largemouth bass stayed longer in the vegetation be resource limited (Werner 1986; Persson et al. than had been predicted, also pointing to predation 1999). For example, in bluegill populations, pre- risk as a possible cause of the habitat distribution dation risk from largemouth bass forces small of fish (see Chapters 11–13, this volume). bluegill sunfish to stay in the vegetated littoral zone, where they experience reduced encounter 15.3.2 Size-dependent rates with their preferred prey and, consequently, competitive interactions reduced individual growth rates (Mittelbach and Chesson 1987; Werner et al. 1983b; Mittelbach, The study of Mittelbach (1981) clearly pointed out Chapter 11, this volume). Food and habitat niche the importance of size for ecological interactions overlap between small and large bluegill is there- in fish communities. Considering competitive in- fore very low but, despite this, small and large teractions first, the classical study of Alm (1946) bluegill sunfish will affect each other through on Eurasian perch had already showed that interac- their coupled life cycle (Mittelbach and Chesson tions between size cohorts could lead to severe re- 1987). Because bluegills in the presence of pisci- source limitation resulting in stunted growth in vores use two different habitats during ontogeny this species. Persson (1983) showed that stunting and thereby have the potential to reduce resource in perch was not restricted to the low-productivity levels in both habitats (Mittelbach 1988), negative forest systems studied by Alm but also occurred in indirect interactions between the resources of highly productive systems. It was suggested that bluegill in these habitats are also present. For ex- low habitat heterogeneity of the system promoted ample, an increase in the food resources of juve- slow-growing individuals, leading to stunted niles translates into increased numbers of adults populations. In an enclosure experiment, Persson and increased consumption rates on the adult’s (1987b) also showed that smaller perch were supe- food resources through higher juvenile growth rior to larger perch in competition for pelagic food rates and survival. This leads ultimately to a resources due to the former’s lower metabolic de- decline in the adult’s resources. Conversely, an mands. Recent studies on very small stages of fish increase in food resources for adults increases juve- including larvae have underlined the influence of nile abundance through increased adult fecundity, size-dependent foraging rates and metabolic de- so leading to a decrease in abundance of juvenile mands on the outcome of competitive interactions food resources (Mittelbach and Chesson 1987; (Byström and Garcià-Berthóu 1999). Although it Mittelbach and Osenberg 1993). The crucial point has been shown that roach in general is a superior here is that a life-history stage not limited by the Community Ecology of Freshwater Fishes 327 resources it needs but by recruitment from other teractions were illustrated by the invasion of the life stages will not be able to take full advantage of redside shiner (Richardsonius balteatus) into Paul an increase in its own resource. The abundances Lake (British Columbia). Up to 1950, this lake was and growth rates of juvenile and adult fish in a inhabited only by rainbow trout (Salmo gairdneri) population are ultimately determined by the pro- (Johannes and Larkin 1961), where all life stages ductivities of resources available to them and, as a fed on a mixture of plankton, benthos and terrestri- consequence, by the relative sizes of the littoral al insects. After the invasion, adult trout started to and pelagic habitats. Littoral macroinvertebrates feed on the redside shiner and dropped plankton are most often the resource that limits bluegill from their diet. As a result, their growth rates in- populations as a whole, resulting in negatively creased. However, competition from redside shin- density-dependent growth rates in juveniles but ers decreased the growth rate of juvenile trout, positively density-dependent growth rates in which negatively affected the recruitment of small adults (Mittelbach and Osenberg 1993). This topic trout to piscivorous size classes (Johannes and also has a bearing on the evolution of life histories Larkin 1961). The different stages of trout were as discussed by Hutchings in Chapter 7, this thus affected by the redside shiner in opposite volume. ways, with a reduction in the trout population In many lakes bluegill sunfish coexist with as an overall effect. The lake trout–redside shiner pumpkinseed sunfish, which leads to further example also illustrates that the extent to which stage-dependent interactions. Adult pumpkinseed growing fish reach predatory stages is heavily in- use littoral snails, which are not used by adult fluenced by different kinds of density-dependent bluegills, but juvenile pumpkinseeds share the interactions during different parts of their ontoge- macrophyte habitat with juvenile bluegill. An in- ny (see also Mittelbach and Chesson 1987; Persson crease in the productivity of pelagic zooplankton 1988; Persson and Greenberg 1990; Persson et al. will lead to an increase in the fecundity of adult 1999). bluegill, which leads to an increase in juvenile bluegill density. This then translates into de- 15.3.4 Ontogenetic constraints creased growth rate and survival of juvenile and trade-offs pumpkinseed. As a result, adult pumpkinseed abundance will decrease and snail abundance in- In fish, as in many other taxa, changes in body size crease (Mittelbach and Chesson 1987; Mittelbach often necessitate a series of ontogenetic niche and Osenberg 1993). shifts toward larger prey sizes and shifts in habitat to allow for an optimal or even positive net energy 15.3.3 Size-dependent intake (Werner and Gilliam 1984; Persson and mixed interactions Greenberg 1990). The presence of such size- dependent ontogenetic niche shifts imposes a se- Stage- or size-structured dynamics prevalent in ries of constraints on the organism. This is because fish populations also mean that the nature of the traits are expressed over several ontogenetic stages interactions will change over ontogeny (Werner of an individual and are often highly intercorrelated and Gilliam 1984; Werner 1986; Persson 1988; because of pleiotropic mutations (ontogenetic co- Mittelbach and Osenberg 1993; Olson et al. 1995). variance) (Werner 1988). As natural selection oper- Changes in habitat and food resources used over ates on morphological and behavioural traits over ontogeny cause top predators to be exposed to both the whole life cycle, traits that are optimal in one competitive and predatory interactions over their ontogenetic niche may be suboptimal in others life (Persson 1988; Mittelbach and Osenberg 1993; (Werner and Hall 1979; Werner 1986, 1988; Persson Olson et al. 1995; Mittelbach, Chapter 11, this vol- 1988). In such cases, genetic covariances will set ume). Some species can be both predator and prey. profound constraints on how efficient an individual The community consequences of these mixed in- can be at each of these ontogenetic stages, which 328 Chapter 15 typically results in ontogenetic trade-offs (Persson vores being better competitors than the juvenile 1988; Werner 1988). A typical example of such a stage of the piscivore. This is a result of ontogenet- situation is piscivorous fish that can undergo ic trade-off costs in the piscivore when in their two to three ontogenetic niche shifts over their planktivore phase. Increased abundance of the life cycle. The body designs that are optimal for competitively superior, smaller species will have a planktivory, benthivory and piscivory can be negative effect on the growth and survival of the ju- quite different. For the planktivorous stage, typical venile size classes of the potentially larger species morphological traits are compressed body and a (competitive asymmetry). In contrast, there will small gape size, which allows the fish to capture be a positive effect on the growth and fecundity of relatively small and non-evasive prey items effi- the piscivorous adult size classes and a negative ciently at high swimming speed (Werner 1977). effect of the piscivores on the smaller species Benthivorous feeding, especially in vegetated (predatory asymmetry) (Persson 1988). The out- habitats, requires a relatively low cruising speed, a come of such complex interactions will be deter- relatively deep body, laterally inserted pectoral mined by the relative strengths of predation and fins and enlarged dorsal fins (Eklöv and Persson competition. Several cases have been described 1995). Piscivorous feeding involves large evasive where the competitive asymmetry among juve- prey, which requires traits such as high attack niles is intense, leading to competitive juvenile speed, large gape size and an attacking foraging bottlenecks in the recruitment to piscivorous size mode (Juanes et al., Chapter 12, this volume). Exist- classes (Johannes and Larkin 1961; Persson 1988). ing data suggest that different piscivorous species This type of asymmetric interaction, particularly have solved the problem of ontogenetic trade-offs in changes in the relative strengths of the predatory different ways and that the different solutions have versus competitive advantage, may have major had feedback effects on the impact of different consequences for overall community and eco- species on overall trophic dynamics (Mittelbach system dynamics as will be considered in the and Persson 1998). following. The topic of predatory and competitive Based on the constraints imposed by ontoge- interactions that change with ontogeny is also netic covariance, a general hypothesis has been covered by Mittelbach (Chapter 11, this volume) advanced. This states that a species undergoing but from a viewpoint determined by a theoretical substantial niche shifts during its lifespan will perspective. have a lower maximum efficiency in any of the on- togenetic niches it utilizes than another species undergoing less substantial ontogenetic niche 15.4 FISH COMMUNITY shifts (Werner and Hall 1979; Werner 1986; STRUCTURE, PRODUCTIVITY Persson 1988). For example, a juvenile of a species AND HABITAT STRUCTURE that is piscivorous as an adult will be an inferior planktivore feeder compared with a species spe- 15.4.1 Size-structured community cializing as a planktivore. Experimental support patterns and productivity for this hypothesis has been provided for at least two species constellations: the perch–roach inter- Productivity is a major factor affecting trophic action and the largemouth bass–bluegill interac- dynamics in ecological systems. Studies of fish tion (Werner 1977; Persson 1988). communities in temperate lakes of different pro- As a consequence of ontogenetic trade-offs in ductivities have shown that substantial changes to piscivores, Persson (1988) suggested that the inter- the fish fauna take place along the productivity actions between piscivores and planktivores are gradient (Leach et al. 1977; Persson et al. 1991). characterized by a high degree of asymmetry. By This shift involves a change from a numerical definition, piscivores have a predatory advantage dominance of Salmoniformes in temperate sys- but this is counteracted by the juveniles of plankti- tems of low productivity to a dominance of percids Community Ecology of Freshwater Fishes 329

in lakes of medium productivity. In highly produc- Chlorophyll tive lakes, the fish community is dominated by 80 Benthic cyprinids. Studies on the relationship between Pelagic productivity and fish community structure have mainly focused on European systems. Even so, 60 studies of North American systems suggest a simi- lar shift in fish community structure with the ex- 40 ception that centrarchids, not present in Eurasian systems, increase steadily in abundance with Piscivorous perch or productivity as do cyprinids (Colby et al. 1972; phytoplankton biomass 20 Leach et al. 1977; Oglesby et al. 1987). While shifts in fish communities along the productivity gradient have been recognized for a 0 0.01 0.1 1 long time, the recognition of the implications of Phosphorus loading (g m–2) size-structured processes for these shifts and the trophic dynamical consequences is more recent Fig. 15.3 Changes in the percentages of pelagic and (Persson 1988; Persson et al. 1991, 1992). The im- benthic piscivorous perch biomasses of total pelagic plications of the shifts in fish communities along fish biomass and phytoplankton biomass (chlorophyll the productivity gradient for the dynamics of a, mgl-1) along a phosphorus loading gradient in Swedish trophic relations are a result of the high proportion lakes. (Source: data from Persson et al. 1991, 1992; of piscivores in the biomass of fish in systems Persson 1994.) dominated by percids. The proportion of pis- civores is much lower in systems dominated by Greenberg 1990). Correspondingly, it has been Salmoniformes or cyprinids (Persson et al. 1991, shown that roach are superior zooplankton feeders 1992; Persson 1994). In percid-dominated com- to perch (Persson 1988). This concerns both the munities, the proportion of piscivores, mainly capacity to depress zooplankton densities and the made up of piscivorous perch, forms up to 80% of capacity to feed on smaller zooplankton forms, total fish biomass (Fig. 15.3). In contrast, the pro- which are a result of intense predation pressure portion of the total fish biomass formed by pisci- (Persson 1994). For example, the attack rates of vores in highly productive systems is low (£20%) roach when feeding on Daphnia and cyclopoid (Persson et al. 1991, 1992). Qualitatively, benthic copepods are substantially higher than those of and pelagic habitats show the same pattern of perch (Persson 1988). It has also been experi- change in the importance of piscivores (Fig. 15.3). mentally shown in both pond and whole-lake However, differences are present in that while pis- experiments that roach, by depressing resource civores are present in low numbers in the benthic levels of pelagic zooplankton, force juvenile perch habitat of very low productivity lakes, piscivores to shift their feeding to benthic macroinverte- may be totally absent in the pelagic habitat of these brates. This leads to a competitive juvenile bottle- lakes (Persson et al. 1992). neck in the recruitment of perch to piscivorous The change in fish community structure can stages (Persson and Greenberg 1990; Persson et al. be set in the general conceptual framework of on- 1999). togenetic trade-offs and asymmetric interactions In highly productive systems, this competitive discussed in the previous section. The shift in a juvenile bottleneck leads to perch populations numerical dominance from percids to cyprinids being dominated by small non-piscivorous size with increasing productivity has been related to classes (Persson 1988; Persson and Greenberg the ability of roach to outcompete the adult pisciv- 1990). In contrast, perch populations in moder- orous perch at the juvenile stage when they are ately productive systems are generally character- feeding on zooplankton (Persson 1988; Persson and ized by having a greater proportion of large-sized 330 Chapter 15 individuals. This change in the size structure of of juvenile bluegill on YOY bass is larger than the perch populations from highly productive lakes to reverse (Olson et al. 1995). moderately productive lakes has been related to at In contrast to the perch–roach system, the least two factors: (i) a decrease in the abundance of abundance relationship between bluegill and blue-green algae eaten by roach but not by perch largemouth bass is always positive. This difference with decreasing productivity; and (ii) a relative in density relationship between perch–roach and increase in the littoral zone and an increase in bass–bluegill can be related to differences in the structurally complex habitats such as submerged life histories of the two piscivores (Olson et al. vegetation, which favours perch (Persson 1988, 1995). Bass become piscivorous in their first year, 1993). Both of these factors provide possible mech- whereas it may take several years for perch to reach anisms by which the competitive advantage of piscivorous stages. The potential for competing cyprinids may decrease, allowing a larger fraction prey to affect perch recruitment to piscivorous of perch to reach larger size classes. An increased stages is thus higher. A functional basis for the dif- number of perch recruiting to larger size classes ferences in life histories has been demonstrated by leads to the possibility that perch predation may showing that largemouth bass have a larger gape affect the density of roach and other cyprinids width for a specific body length than do Eurasian (Persson 1988). The presence of a predatory impact perch (Mittelbach and Persson 1998). Further- by perch on roach is supported by the fact that the more, gape width is negatively related to size at median size of roach populations is higher in lakes switch to piscivory and positively related to size at where the component of total fish biomass con- age 1, suggesting a coupling between morphology sisting of piscivorous perch is higher (Persson et al. and ecological performance. Finally, largemouth 1991). The complex size-dependent interactions bass have shorter prey manipulation times on fish between roach and perch resulting from both pre- prey than perch (Mittelbach and Persson 1988). dation and competition should be kept in mind In conclusion, the studies on the perch–roach and when reading the chapters by Shepherd and Pope the Centrarchidae systems suggest that size- (Chapters 7 and 8, Volume 2), Schnute and structured interactions are of overwhelming Richards (Chapter 6, Volume 2) and Pauly and importance in structuring fish communities and Christensen (Chapter 10, Volume 2). that the relationship between top predators and their competing prey may take different forms 15.4.2 Impact of piscivore depending on piscivore life histories. life history 15.4.3 Habitat structure Depending on productivity, the abundance rela- tionship between perch and its prey and competi- Habitat structure, often in the form of submerged tors may be either positive or negative (Fig. 15.3). vegetation, affects both competitive and predatory Similar studies on the Centrarchidae system have interactions (Winfield 1986; Savino and Stein 1989; shown that the bluegill sunfish makes up the ma- Persson 1993; Persson and Crowder 1998). Besides jority of total fish biomass in many lakes in central an overall reduction in predation efficiency of pisci- North America (Mittelbach and Osenberg 1993). vores, complex habitats like vegetation may affect Juvenile bluegill may have a negative effect on the the relationship between body size and predation growth of its competitors, including the juveniles mortality of prey fish (Savino and Stein 1989; of the major piscivorous predator, largemouth bass Persson and Eklöv 1995; Persson and Crowder (Mittelbach 1988; Olson et al. 1995), via the same 1998). The mechanisms behind the decreased pre- mechanism as outlined above for European perch dation efficiency in complex habitats can be both and roach. Correspondlingly, it has been experi- decreased encounter rate between predator and mentally demonstrated that the per-capita effect prey and decreased capture success of the predator Community Ecology of Freshwater Fishes 331 once the prey has been encountered (Savino and zooplankton. Subsequently, a large number of pa- Stein 1989; Juanes et al., Chapter 12, this volume). pers have demonstrated the size-selective effects of Habitat structure in the form of vegetation may also planktivory on zooplankton communities. More- reverse the outcome of competitive interactions over, studies by Hrbàcek et al. (1961), Henrikson among refuging prey fish. For example, the compet- et al. (1980) and other European researchers itive advantage of juvenile roach over juvenile showed that the effect of fish predation was not perch in open water was reversed in refuges in vege- restricted to an effect on zooplankton but also trans- tation because of the effects of structure per se on formed into effects on phytoplankton. Studies by roach foraging performance and because vegetation researchers including Shapiro and Wright (1984), structure harbours prey that juvenile perch are bet- Carpenter et al. (1987) and Carpenter and Kitchell ter at foraging than are juvenile roach (Winfield (1993) subsequently demonstrated similar strong 1986; Persson 1991, 1993). top-down effects of piscivorous and planktivorous Because habitat structure affects both predator– fish on phytoplankton communities and nutrient prey and competitive relationships, habitat dynamics in North American systems. More long- structure is also expected to affect community term studies carried out by Mittelbach et al. (1995) patterns (see Mittelbach and Osenberg 1993 showed strong effects over several trophic linkages, and bluegill example above). Although the major covering piscivores to phytoplankton, following change in fish communities of European systems a winter-kill affecting largemouth bass. A most are associated with changes in productivity, it has dramatic example of the impact of a piscivorous been hypothesized that this change is more cau- top predator is also provided by the introduction sally linked to changes in habitat structure than to of the Nile perch in Lake Victoria (Kaufman productivity itself (Persson et al. 1992; Persson 1992; Pitcher and Hart 1995; Kitchell et al. 1997). 1994). This hypothesis relates to the observation In this lake, the biomass of the native fish that the importance of submerged vegetation is fauna, particularly Haplochromis species, have generally at a maximum in moderately productive decreased to very low levels and many species have lakes (Wetzel 1979) where piscivore biomass is eventually gone extinct (Fig. 15.4). The effects of also at a maximum. Juvenile perch and roach prefer to forage in the vegetation if the risk of predation is high in the open water, and a high abundance of pelagic piscivores in lakes of intermediate produc- 80 tivity may thereby reverse the outcome of compet- 70 itive interactions between juveniles in favour of Native species

) Introduced species 60 perch (Winfield 1986; Persson 1993; Persson and –1 Eklöv 1995). 50 40 15.5 EFFECTS OF FISH 30 ON LOWER TROPHIC 20 Fish biomass (kg ha COMPONENTS 10 0 15.5.1 Fish predation and 1970 1975 1977 1983 1993 trophic dynamics Year

The classic paper by Brooks and Dodson (1965) Fig. 15.4 Changes in standing stocks of native versus showed that predation from planktivorous fish introduced fishes in Lake Victoria between 1970 and decreased both the abundance and mean size of 1993. (Source: data from Kaufman 1992.) 332 Chapter 15 these dramatic changes in fish fauna on lower induced indirect effect has been demonstrated trophic components are difficult to discern as the by piscivorous largemouth bass in pond experi- lake has undergone substantial eutrophication ments. Piscivore-mediated habitat choice in small over the same time period. Comparative studies on bluegills caused an increase in zooplankton abun- lakes have provided evidence for the effects of dance in the open water compared with controls top predators on lower trophic components. The lacking largemouth bass (Turner and Mittelbach changes in the importance of piscivores along the 1990). productivity gradient of Swedish lakes also in- volved feedback effects on other trophic levels 15.5.2 Effects of fish on such as zooplankton and phytoplankton (Fig. 15.3) nutrient cycling (Persson et al. 1992; Persson 1994). For example, a ten-fold increase in phosphorus loading from 0.03 In addition to direct mortality and behavioural to 0.3gm-2 year-1 only led to a minor increase shifts in prey caused by fish predators, fish also af- in phytoplankton biomass, suggesting that an in- fect trophic dynamics through their effects on nu- creased piscivore biomass may prevent an in- trient cycling. In many systems, biota including creased phosphorus loading leading to an increase fish represent major storage pools of nutrients that in phytoplankton biomass within this range of limit productivity, and consumer–resource inter- phosphorus loadings. In contrast, phytoplankton actions greatly affect nutrient cycling in different biomass increased steadily with phosphorus components of the food web. For example, Kaiser loading in highly productive systems with a low and Jennings (Chapter 16, Volume 2) review the proportion of piscivores (Fig. 15.3). importance of Pacific salmon (Oncorhynchus spp.) Many of the manipulations of fish populations for movements of nutrients between marine and have thus resulted in effects on planktivores freshwater habitats. It may be important to con- and zooplankton to phytoplankton. Traditionally, sider nutrient recycling explicitly when the ratio these have been interpreted as direct effects of of different essential resources, such as carbon, predator consumption. However, behavioural phosphorus and nitrogen, is different in ingested studies of predator–prey interactions have also em- food compared with egested food (Sterner 1990). phasized non-lethal indirect effects of predators on For example, when predation pressure from plank- community dynamics, such as effects on prey fish tivorous fish on zooplankton is low, Daphnia activity and habitat use (Power 1987; Turner and dominate and they have a lower N:P ratio than do Mittelbach 1990). Power (1987) showed striking copepods, which often dominate when predation differences in the standing stock of algae between pressure from planktivorous fish on zooplankton shallow margins and deeper central areas within is high (Sterner and Hessen 1994). Such differences pools of a Panama stream related to the distribu- in nutrient signatures can be used to trace links tion of the armoured catfish (Loricariidae). Data between members of a trophic web (Polunin and suggested that although large catfish were severely Pinnegar, Chapter 14, this volume). Nutrient recy- resource limited in the deeper parts of pools, they cling and transformation by zooplankton is widely avoided shallow areas because of the greater risk of recognized as an important source of nutrients for predation from birds (Power 1987). In contrast, phytoplankton production (Vanni 1996). Experi- small catfish were much more abundant in mental studies have also documented that direct shallow water in order to avoid predatory fishes nutrient recycling by fish will affect phytoplank- occupying deeper waters, but did not deplete algal ton community structure. Vanni et al. (1997) resources there (Power 1987). Altogether, the evi- showed that in lakes dominated by planktivores, dence suggested that risk-sensitive foraging by fish effects on algal community composition can large catfish resulted in a behaviourally induced exceed the effects of nutrient recycling by zoo- indirect interaction between avian predators and plankton. As fish are rarely limited by phosphorus benthic algae. Another example of a behaviourally availability but are by energy, the N:P ratio ex- Community Ecology of Freshwater Fishes 333 creted from fish is generally low, which stimulates In correspondence with the observation that phytoplankton growth (Schindler and Eby 1997). fish may translocate nutrients from the littoral/ The effects of size structure on nutrient recycling benthic habitats to pelagic habitats, recent biogeo- by fish will also be present because the ratio of chemical analyses of phosphorus cycles in the phosphorus-rich bone mass to total body mass pelagic zones of lakes suggest that observed levels varies with fish size (Vanni 1996). However, the of primary productivity are too high to be sup- effect of fish body size on nutrient recycling is ported by pelagic recycling alone. For example, complex as a result of the interaction between one-third of the pelagic primary production of physiological and ecological processes (Schindler Mirror Lake (New Hampshire, USA) has been esti- and Eby 1997). The latter results from ontogenetic mated to be supported by phosphorus coming from niche shifts, where larger fish generally eat larger elsewhere (Caraco et al. 1992). Correspondingly, prey with high phosphorus concentrations while both Vanni (1996) and Schindler et al. (1996) argue smaller fish eat small prey with a generally lower strongly in recent reviews that pelagic nutrient phosphorus concentration. budgets can only be balanced by considering If the effects of nutrient translocation between biologically driven phosphorus from the littoral. habitats due to benthivory are also considered, the The current knowledge therefore suggests that the impact of fish will become even larger (Schindler littoral zone is a source of nutrients for pelagic food et al. 1996). It has for a long time been recognized webs and that the pelagic habitat consequently is a that littoral habitats are linked biogeochemically sink for littoral productivity in most lakes. to the open waters of lakes (Wetzel 1979). Because invertebrates in littoral zones generally have a 15.5.3 Alternative states in highly higher biomass, production and diversity than productive systems those in open water, littoral zones form a potential resource base for pelagic production. Fish have a The interaction between vegetation structure substantial impact on lake nutrient cycling be- and trophic structure and the different pathways cause they function as vectors for transport of nu- through which fish may affect nutrient recycling trients from littoral and benthic habitats through has led to the proposal that there exists a potential their feeding on littoral benthic prey (Schindler for alternative stable states in highly productive et al. 1996). Fish that migrate onshore–offshore systems (Scheffer 1990, 1998; Blindow et al. 1993). may thus increase nutrient translocation above One state is characterized by high phytoplankton that accounted for by biogeochemical processes biomass, low biomass of zooplankton and high alone (Schindler et al. 1996). Under some circum- biomasss of planktivore/benthivore fish; the stances, this nutrient translocation caused by fish other state is characterized by high biomass of may exceed the external loading of nutrients macrophytes, low biomass of phytoplankton, high (Brabrand et al. 1990). The translocation of biomass of zooplankton and a high biomass of nutrients may also involve nutrient release from piscivores. Theoretical studies have suggested sediments (Brabrand et al. 1990; Vanni 1996). that the existence of the two alternative states Finally, translocation of nutrients may occur should be restricted to shallow highly productive over large distances, exemplified by anadromous lakes, whereas in deep highly productive lakes sockeye salmon (Oncorhynchus nerka) (Kaiser only the state with high biomass of phytoplankton and Jennings, Chapter 16, Volume 2). The salmon should be present (Scheffer 1990, 1998). Corre- were a major vector for the transport of nutrients spondingly, field evidence that highly productive from marine systems to nutrient-poor freshwater lakes may be dominated by macrophyte produc- systems, and human-caused inhibition of their tion is restricted to shallow lakes (Blindow et al. upwards migration into fresh waters resulted in 1993; Scheffer 1998). Nevertheless, the evidence major decreases in the productivity of these for the presence of alternative stable states in systems (Stockner and MacIsaac 1996). shallow highly productive lakes provides a nice 334 Chapter 15 illustration of the impact of fish on lake ecosystem developed theoretical framework, physiologically dynamics. structured models have been successfully applied to the study of cohort interactions and cannibal- ism in fish (Persson et al. 1998; Claessen et al. 15.6 FROM INDIVIDUAL- 2000). A different perspective on individual-based LEVEL PROCESSES TO models that predict population processes is given POPULATION DYNAMICS by Huse et al. (Chapter 11, Volume 2). Considering intercohort competition first, the So far individual-level capacities, like the abilities foraging rate as a function of body weight is to forage and avoid predators, have been connected commonly described by a power function (Werner to community patterns, such as how trade-offs be- 1988). For a given prey size, the attack rate a(w) as a tween maximizing energy gains and minimizing function of weight (w) has been found to be a predation risk may affect the size distribution of hump-shaped function of predator size (Byström size cohorts and species (Mittelbach 1981; Werner and Garcìa-Berthóu 1999). The initial increase in 1986; Osenberg et al. 1988; Persson 1988). Essen- the foraging capacity with consumer size is due to tially, community patterns have thus been pre- an increase in visual acuity and locomotor ability, dicted from individual, size-dependent, capacities. both of which will affect the encounter with prey Although this approach has been relatively suc- (Persson et al. 1998). The decreasing part of the cessful in increasing our understanding of com- function is due to a decrease in retinal rod density munity patterns, a major impediment has been and hence the capacity to discern small prey (Breck that the population dynamics between interacting and Gitter 1983). Furthermore, attack efficiency species has been neglected. Stage-based popula- may also decrease above a given body size due to tion models based on just two life stages, juveniles decreased ability to make fine-tuned manoeuvres and adults, have been shown to successfully pre- (Persson 1987b). A function that encapsulates dict the abundance patterns of different stages and these aspects is the following formula: species in some cases (Mittelbach and Chesson a 1987; Mittelbach and Osenberg 1993). At the same È w Ê w ˆ˘ aw()=- AÍ expÁ 1 ˜˙ (15.1) time, the fact that fish grow during most of their ÎwooË w ¯˚ lives and that individual growth is strongly limited by food availability (Jobling, Chapter 5, this vol- where A is the maximum attack rate, wo is the ume) calls for an approach that explicitly handles body size at which the maximum rate is achieved these features. and a is a size scaling exponent (Fig. 15.5) (Persson Physiologically structured models represent an et al. 1998). Empirical support for the form of this approach that explicitly links the dynamics that function is found in Byström and Gàrcia-Berthóu takes place at the level of population to the behav- (1999). Using the same principle, the function iour of individual organisms (De Roos 1997). These for handling time as a function of body size can be models are based on a state concept at each of two derived (Persson et al. 1998; Claessen et al. 2000). levels of organization: an i-state, which represents The metabolic rate of fish also increases with the state of the individual in terms of a collection body size and for many fish species a size scaling of characteristic physiological traits such as size, exponent (a) around 0.75 has been found (Persson age, sex and energy reserves, and a p-state, which is et al. 1998; Claessen et al. 2000). Taking both forag- the frequency distribution over the space of pos- ing gains, expressed as the attack rate times energy sible i-states (De Roos 1997). The core of physio- content per prey and assimilation efiiciency, and logically structured models is the description of metabolic costs associated with gaining this the individual state. The population state becomes energy into consideration allow us to derive the a matter of book-keeping of the individuals in dif- minimum level of the resource that will allow the ferent i-states. Although representing a recently fish to maintain itself. For planktivorous prey, Community Ecology of Freshwater Fishes 335

(a)3 (b) 2

1

) 1 –1 ) –1

0.1 Attack rate (l s Minimum resource density (l

0.01 0.1 0.001 0.01 0.1 1 10 100 0.001 0.01 0.1 1 10 100 Body mass (g) Body mass (g)

Fig. 15.5 (a) Attack rate and (b) minimum resource requirements of a planktivorous fish as a function of fish size. -1 Parameter values: A = 2.3ls , wo = 17.4g, a=0.5 and zooplankton size = 0.6mm. (Source: see Persson et al. 1998 for derivation.)

where the weight exponent in equation 15.1 has Persson 1986; Auvinen 1994). The cycle length in been shown to vary between 0.59 and 0.67 these populations varies between 2 and 5 years (Mittelbach 1981; Byström and Garcìa-Berthóu (Aass 1972; Hamrin and Persson 1986) (Fig. 15.6) 1999), the minimum resource level is a monotoni- and is probably driven by variation in recruitment cally increasing function of body size (Fig. 15.5). caused by the adults suffering severe mortality due The interpretation of this is that smaller-sized to starvation. Roach populations have been shown individuals will have the lowest minimum re- to demonstrate similar oscillations (Townsend source demand level and hence be competitively et al. 1990). superior to larger individuals with respect to So far, I have only considered a size-structured exploitative competition. population and a non-structured population of A general result derived from this model is that zooplankton. In fish populations, both predator consumer–resource interactions, for example the and prey are generally size-structured, which par- interaction between planktivorous fish and zoo- ticularly concerns the interaction between pisciv- plankton, are likely to result in population oscilla- orous predators and their fish prey (Hambright tions. These oscillations are driven by recruiting 1994; Christensen 1996). The vulnerability of fish individuals that by their pulsed appearance larval fish to raptorial predators has been found to will heavily depress the zooplankton resource and increase to a maximum and thereafter decrease thereby reduce fecundity and increase starvation as prey fish size increases. An important conse- risk in larger mature fish. There are a number of quence of such a form of the attack rate is that examples of planktivorous fish that support this newborn prey fish, depending on the size distribu- modelling result. A classic example is population tion of potential predators, may subject these oscillations in zooplanktivorous cisco (vendace) predators to competition for a shared resource (Coregonus albula) (Aass 1972; Hamrin and such as zooplankton before the prey fish are effi- 336 Chapter 15 ciently captured by their predators. With a further petition for a basic resource suggested that the increase in size, prey fish vulnerability will in- dynamics can be characterized by a mixture of crease with size to a peak and thereafter decrease. high-amplitude recruiter-driven dynamics inter- As it is often the case that both piscivores and prey spersed with periods of cannibal-driven dynamics fish increase in size over time, the time span over (Claessen et al. 2000; Juanes et al., Chapter 12, this which a prey fish cohort is susceptible to predation volume). The amplitude of cannibal-driven dy- will depend on the growth rates of both predator namics varies from high-amplitude dynamics to and prey. stable equilbrium (Claessen et al. in press). Inter- Investigations of a physiologically structured estingly, cannibalistic individuals with accelerat- model including both cannibalism and com- ing growth rates (Fig. 15.7) may appear under several conditions given that the cannibals feed on relatively large victims. An empirical study of the population dynamics of an allopatric cannibalistic 15 Cisco perch population showed that the dynamics of the perch population could be separated into two Cladocerans phases (Persson et al. 2000). In one phase, adult perch dominated and recruitment of YOY perch to 10 1-year-old perch was low due to high cannibalism on young and small stages of YOY perch. In these years, adult perch had an asymptotic length of Abundance 5 around 180mm. This phase was interrupted by a phase when the number of adult perch decreased possibly due to starvation, leading to a period of strong recruitment of 1-year-old perch. The 0 growth of the few surviving adult perch acceler- 1970 1971 1972 1973 1974 1975 1976 1977 ated as a result of the energy that older and thus Year larger YOY perch represented, leading to a ‘double asymptotic’ growth curve (Fig. 15.7). Interestingly, Fig. 15.6 Abundances of young-of-the-year cisco the growth patterns of perch in the two different (Coregonus albula) (catch per unit effort, g ¥ 10-2) and cladoceran zooplankton (no. per litre) in Lake Bolmen phases were the same as the growth patterns pre- in the years 1970–77. (Source: data from Hamrin and dicted by the model (Fig. 15.7) (Claessen et al. Persson 1986.) 2000).

Fig. 15.7 Growth rates of (a) (b) cannibalistic perch in years when 250 250 Data young-of-the-year (YOY) perch were cannibalized (a) early and (b) 200 Model 200 in years when YOY perch survived for a longer period in Lake 150 150 Abborrtjärn 3 (squares, solid lines) and as predicted by the model 100 100 (circles, dotted lines). In (b), YOY Length (mm) perch surviving for a longer period 50 50 were present from the time when cannibalistic perch were 4 years 0 0 0123456 0123456 old. (Source: data from Claessen et al. 2000; Persson et al. 2000.) Age (years) Age (years) Community Ecology of Freshwater Fishes 337

populations with special reference to perch. Report 15.7 CONCLUSIONS from the Institute of Freshwater Research, Drottning- holm No. 25. The analyses of the dynamics of fish populations Appelberg, M. (1998) Restructuring of fish assemblages using physiologically structured population mod- in Swedish lakes following amelioration of acid stress els are so far restricted to intercohort competition through liming. Restoration Ecology 6, 343–53. Auvinen, H. (1994) Intra- and interspecific factors in the and cannibalism. The results of these analyses dynamics of vendace (Coregonus albula (L.)) popula- are encouraging and suggest that physiologically tions. PhD thesis, University of Helsinki, Finland. structured models are a useful tool for increasing Baltz, D.M., Moyle, P.M. and Knight, N.J. (1982) Com- our understanding of the dynamics of fish popula- petitive interactions between benthic stream fishes, tions. The analyses of intercohort competition and riffle sculpin, Cottys gulosus, and speckled dace, cannibalism also suggest that there is potential to Rhinichthys osculus. Canadian Journal of Fisheries and Aquatic Sciences 39, 1502–11. reach a substantial degree of generality in results. Bergman, E. (1988) Foraging abilities and niche breadths Is it possible to go further to more complicated of two percids, Perca fluviatilis and Gymnocepahlus scenarios, involving, say, two size-structured fish cernua, under different environmental conditions populations of which one is a predator on the other Journal of Animal Ecology 57, 443–53. species, and continue to have a substantial degree Blindow, I., Andersson, G., Hargeby, A. and Johansson, of generality in model predictions? Examples S. (1993) Long-term pattern of alternative states in might be perch–roach and largemouth bass– two shallow eutrophic lakes. Freshwater Biology 30, 159–67. bluegill interactions. If this is the case, physiologi- Brabrand, Å., Faafeng, B.A. and Nilssen, J.P.M. (1990) cally structured population models may turn out Relative importance of phosphorous supply to phyto- to be a valuable approach for the analysis of fish plankton production: fish excretion versus external community dynamics, including the impact of loading. Canadian Journal of Aquatic and Fisheries fish on their resources as discussed in this chapter. Sciences 47, 364–72. Ongoing research suggests that this may be pos- Breck, J.E. and Gitter, M.J. (1983) Effect of fish size on the sible. The dynamics of more complex configura- reactive distance of bluegill (Lepomis macrochirus) sunfish. Canadian Journal of Fisheries and Aquatic tions appear to be largely driven by combinations Sciences 40, 162–7. of the mechanisms analysed in the simpler con- Brett, J.R. and Groves, T.D.D. (1979) Physiological ener- figurations, such as recruit-driven and cannibal- getics. In: W.S. Hoar, D.J. Randall and J.R. Brett (eds) driven cycles. These analyses and the analysis of Fish Physiology, Vol. VIII. London: Academic Press, Claessen et al. (2000) also suggest that important pp. 280–344. elements of the dynamics observed in both the Brooks, J.L. and Dodson, S.I. (1965) Predation, body size and composition of plankton. Science 150, 28–35. models and empirical systems are not accessible to Byström, P. and García-Berthóu, E. (1999) Density depen- analysis outside the domain of physiologically dent growth and stage-specific competitive interac- structured population models. Thus, to substan- tions in young fish. Oikos 86, 217–32. tially increase our theoretical understanding of Caraco, N.F., Cole, J.J. and Likens, G.E. (1992) New the implications of size-structured interactions for and recycled primary production in an oligotrophic fish community dynamics, physiologically struc- lakes: insights for summer phosphorus dynamics. tured population models may be indispensable. Limnology and Oceanography 37, 590–602. Carpenter, S.R. and Kitchell, J.F. (eds) (1993) The Trophic Cascade in Lakes. Cambridge: Cambridge University Press. REFERENCES Carpenter, S.R., Kitchell, J.F., Hodgson, J.R., et al. (1987) Regulation of lake primary productivity by food web Aass, P. (1972) Age determination and year class fluctua- structure. Ecology 68, 1863–76. tions of ciscoe (Coregonus albula) in the Mjösa hy- Christensen, B. (1996) Predator foraging capabilities and droelectric reservoire. Report from the Institute of prey antipredator behaviours: pre- versus postcapture Freshwater Research, Drottningholm 52, 4–21. constraints on size-dependent predator–prey interac- Alm, G. (1946) Reasons for the occurrence of stunted fish tions. Oikos 76, 368–80. 338 Chapter 15

Claessen, D., de Roos, A.M. and Persson, L. (2000) Kitchell, J.F., Schindler, D.E., Ogutu-Ohwayo, R. and Dwarfs and giants: cannibalism and competition in Reinthal, P.M. (1997) The Nile perch in Lake Victoria: size-structured populations. American Naturalist interactions between predation and fisheries. Ecologi- 155, 219–37. cal Applications 7, 653–64. Colby, P.J., Spangler, G.R., Hurley, D.A. and McCombie, Krebs, C.J. (1972) Ecology. New York: Harper & Row. A.M. (1972) Effects of eutrophication on salmonid Leach, J.H., Johnson, M.G., Kelso, J.R.M., Hartmann, J., communities in oligotrophic lakes. Journal of the Nümann, W. and Entz, B. (1977) Responses of percid Fisheries Research Board of Canada 29, 1602–12. fishes and their habitats to eutrophication. Journal of De Roos, A.M. (1997) A gentle introduction to physio- the Fisheries Research Board of Canada 34, 1964–71. logically structured population models. In: S. MacArthur, E.R. and Wilson, E.O. (1967) The Theory Tuljapurkar and H. Caswell (eds) Structured of Island Biogeography. Princeton, NJ: Princeton Population Models in Marine, Terrestrial and University Press. Freshwater Systems. New York: Chapman & Hall, Magnuson, J.J., Crowder, L.B. and Medvick, P.A. (1979) pp. 119–204. Temperature as an ecological resource. American Eklöv, P. and Persson, L. (1995) Species-specific anti- Zoologist 19, 331–43. predator capacities and prey refuges: interactions be- Magnuson, J.J., Tonn, W.M., Banerje, A., Toivonen, J., tween piscivorous perch (Perca fluviatilis) and juvenile Sanchez, O. and Rask, M. (1998) Isolation vs. extinc- perch and roach (Rutilus rutilus). Behavioral Ecology tion in the assembly of fishes in small northern lakes. and Sociobiology 37, 169–78. Ecology 79, 2941–56. Elliott, J.M. (1975) The growth rate of brown trout Salmo Mittelbach, G.G. (1981) Foraging efficiency and body trutta L. fed on maximum rations. Journal of Animal size: a study of optimal diet and habitat used in bluegill Ecology 44, 805–21. sunfish. Ecology 62, 1370–86. Hambright, K.D. (1994) Morphological constraints in Mittelbach, G.G. (1988) Competition among refuging the piscivore–planktivore interaction: implications sunfishes and effects of fish density on littoral zone for the trophic cascade hypothesis. Limnology and invertebrates. Ecology 69, 614–23. Oceanography 39, 897–912. Mittelbach, G.G. and Chesson, P.L. (1987) Predation risk: Hammar, J. (1998) Interactive asymmetry and seasonal indirect effects on fish populations. In: W.C. Kerfoot niche shifts in sympatric Arctic char (Salvelinus alpi- and A. Sih (eds) Predation: Direct and Indirect Impacts nus) and brown trout (Salmo trutta): evidence from on Aquatic Communities. Hanover, NH: University winter diet and accumulation of radiocesium. Nordic Press of New England, pp. 537–55. Journal of Freshwater Research 74, 33–64. Mittelbach, G.G. and Osenberg, C.W. (1993) Stage- Hamrin, S.F. and Persson, L. (1986) Asymmetrical structured interactions in bluegill: consequences of competition between age classes as a factor causing adult resource variation. Ecology 74, 2381–94. population oscillations in an obligate planktivorous Mittelbach, G.G. and Persson, L. (1998) The ontogeny of fish. Oikos 47, 223–32. piscivory and its ecological consequences. Canadian Henrikson, L., Nyman, H., Oscarson, H. and Stenson, J. Journal of Fisheries and Aquatic Sciences 55, 1454–65. (1980) Trophic changes, without changes in the exter- Mittelbach, G.G., Turner, A.M., Hall, D.J., Rettig, J.E. nal loading. Hydrobiologia 68, 257–63. and Osenberg, C.W. (1995) Perturbation and resilience. Hrbàcek, J., Dvorakova, M., Korinek, V. and A long term, whole-lake study of predator extinction Prochazkova, L. (1961) Demonstration of the effects of and reintroduction. Ecology 76, 347–60. fish stock on the species composition and the intensity Nilsson, N.-A. (1960) Seasonal fluctuations in the food of the metabolism of the whole plankton association. segregation between trout, char and whitefish in 14 Verhandlungen Internationale Verhandlungen der North-Swedish lakes. Report from the Institute of Internationalen Vereinigung für Theroretische und Freshwater Research Drottningholm 41, 195–205. Angewandte Limnologie 14, 192–5. Nilsson, N.-A. (1963) Interactions between trout and Jobling, M. (1983) Influence of body weight and tempera- char in Scandinavia. Transactions of the American ture on growth rates of Arctic char, Salvelinus alpinus Fisheries Society 92, 276–85. (L.). Journal of the Fish Biology 22, 471–5. Nilsson, N.-A. (1978) The role of size-based predation in Johannes, R.D. and Larkin, P.A. (1961) Competition for competition and interactive segregation in fish. In: S.D. food between redside shiner and rainbow trout in Gerking (ed.) Ecology of Freshwater Fish Production. two British Columbia lakes. Journal of the Fisheries Oxford: Blackwell Scientific Publications, pp. 303–25. Research Board of Canada 18, 203–20. Oglesby, R.T., Leach, J.H. and Forney, J. (1987) Potential Kaufman, L.S. (1992) Catastrophic change in species rich Stizostedion yield as a function of chlorophyll concen- freshwater ecosystems: the lesson from Lake Victoria. tration with special reference to Lake Erie. Canadian Bioscience 42, 846–58. Journal of Fisheries and Aquatic Sciences 44, 166–70. Community Ecology of Freshwater Fishes 339

Olson, M.H., Mittelbach, G.G. and Osenberg, C.W. Persson, L., Diehl, S., Johansson, L., Andersson, G. and (1995) Competition between predator and prey: Hamrin, S.F. (1991) Shifts in fish communities along resource-based mechanisms and implications for the productivity gradient of temperate lakes: patterns stage-structured dynamics. Ecology 76, 1758–71. and the importance of size-structured interactions. Osenberg, C.W., Werner, E.E., Mittelbach, G.G. and Hall, Journal of Fish Biology 38, 281–93. D.J. (1988) Growth patterns in bluegill (Lepomis Persson, L., Diehl, S., Johansson, L., Andersson, G. and macrochirus) and pumpkinseed (L. gibbosus) sunfish: Hamrin, S.F. (1992) Trophic interactions in temperate environmental variation and the importance of onto- lake ecosystems: a test of food chain theory. American genetic niche shifts. Canadian Journal of Fisheries and Naturalist 140, 59–84. Aquatic Sciences 45, 17–26. Persson, L., Leonardsson, K., De Roos, A., Gyllenberg, M. Persson, L. (1983) Food consumption and competi- and Christensen, B. (1998) Ontogenetic scaling of tion between age classes in a perch (Perca fluviatilis) foraging rates and the dynamics of a size-structured population in a shallow eutrophic lake. Oikos 40, consumer–resource model. Theoretical Population 197–207. Biology 54, 270–93. Persson, L. (1986) Temperature-induced shift in foraging Persson, L., Byström, P., Wahlström, E., Hjelm, J. and ability in two fish species, roach (Rutilus rutilus) and Andersson, J. (1999) Interactions among size- perch (Perca fluviatilis): implications for coexistence structured populations in a whole lake experiment: among poikilotherms. Journal of Animal Ecology 55, size-dependent performance and community dynam- 829–39. ics. Oikos 87, 139–56. Persson, L. (1987a) Effects of habitat and season on com- Persson, L., Byström, P. and Wahlström, E. (2000) Canni- petitive interactions between roach (Rutilus rutilus) balism and competition in Eurasian perch: population and perch (Perca fluvuatilis). Oecologia 73, 170–7. dynamics of an ontogenetic omnivore. Ecology 81, Persson, L. (1987b) The effects of resource availability 1058–71. and distribution on resource utilization in perch (Perca Pitcher, T.J. and Hart, P.J.B. (1995) The Impacts of Species fluviatilis). Oikos 48, 148–60. Changes on African Lakes. London; Chapman & Hall. Persson, L. (1988) Asymmetries in competitive and Power, G. (1990) Salmonid communities in Quebec and predatory interactions in fish populations. In: B. Labrador: temperature relations and climate change. Ebenman and L. Persson (eds) Size-structured Polskie Archiwe für Hydrobiologie 37, 13–28. Populations: Ecology and Evolution. Heidelberg: Power, M.E. (1987) Predator avoidance by grazing fishes Springer-Verlag, pp. 203–18. in temperate and tropical streams: importance of Persson, L. (1991) Behavioral response to predators stream depth and prey size. In: W.C. Kerfoot and A. reverses the outcome of competition between prey Sih (eds) Predation: Direct and Indirect Impacts on species. Behavioral Ecology and Sociobiology 28, Aquatic Communities. Hanover, NH: University 101–5. Press of New England, pp. 333–51. Persson, L. (1993) Predator-mediated competition in prey Rahel, F.J. (1984) Factors structuring fish assemblages refuges: the importance of habitat dependent prey along a bog lake successional gradient. Ecology 65, resources. Oikos 68, 12–22. 1276–89. Persson, L. (1994) Natural patterns of shifts in fish com- Savino, J.F. and Stein, R.A. (1989) Behavioral interactions munities: mechanisms and constraints on perturba- between fish predators and their prey: effects of plant tion sustenance. In: I.G. Cowx (ed.) Rehabilitation of density. Animal Behaviour 37, 311–21. Freshwater Fisheries. Oxford: Fishing News Books, pp. Scheffer, M. (1990) Multiplicity of stable states in fresh- 421–34. water systems. Hydrobiologia 200/210, 475–86. Persson, L. and Crowder, L.B. (1998) Fish–habitat interac- Scheffer, M. (1998) Ecology of Shallow Lakes. London: tions mediated via ontogenetic niche shifts. In: Chapman & Hall. E. Jeppesen, Ma. Søndergaard, Mo. Søndergaard Schindler, D.E. and Eby, L.A. (1997) Stochiometry of and K. Christoffersen (eds) The Structuring Role fishes and their prey: implications for nutrient recycl- of Submerged Macrophytes in Lakes. New York: ing. Ecology 78, 1816–31. Springer-Verlag, pp. 3–23. Schindler, D.E., Carpenter, S.R., Cottingham, K.L., He, Persson, L. and Eklöv, P. (1995) Prey refuges affecting X., Hodgson, J.R., Kitchell, J.F. and Soranno, P.A. (1996) interactions between piscivorous perch (Perca Food web structure and littoral zone coupling to fluviatilis) and juvenile perch and roach (Rutilus pelagic trophic cascades. In: G.A. Polis and K.O. rutilus). Ecology 76, 70–81. Winemiller (eds) Food Webs: Integration of Patterns Persson, L. and Greenberg, L.A. (1990) Juvenile competi- and Dynamics. New York: Chapman & Hall, pp. 96– tive bottlenecks: the perch (Perca fluviatilis)–roach 105. (Rutilus rutilus) interaction. Ecology 71, 44–56. Schlosser, I.J. (1987) A conceptual framework for fish 340 Chapter 15

communities in small warmwater streams. In: W.J. Vanni, M.J., Layne, C.D. and Arnott, S.E. (1997) ‘Top- Mathews and D.C. Heins (eds) Community and down’ trophic interactions in lakes: effects of fish on Evolutionary Ecology of North American Stream nutrient recycling. Ecology 78, 1–20. Fishes. Norman, OK: University of Oklahoma Press, Werner, E.E. (1977) Species packing and niche comple- pp. 17–24. mentarity in three sunfishes. American Naturalist 11, Schoener, T.W. (1982) The controversy over interspecific 553–78. competition. American Scientist 70, 586–95. Werner, E.E. (1986) Species interactions in freshwater Schultz, D.C. and Northcote, T.G. (1972) An experimen- fish communities. In: J. Diamond and T.J. Case (eds) tal study of feeding behaviour and interacting coastal Community Ecology. New York: Harper & Row, pp. cutthroat trout (Salmo clarki clarki) and Dolly Varden 344–58. (Salvelinus malma). Journal of the Fisheries Research Werner, E.E. (1988) Size, scaling and the evolution Board of Canada 29, 555–65. of life cycles. In: B. Ebenman and L. Persson (eds) Shapiro, J. and Wright, D.I. (1984) Lake restoration by Size-structured Populations: Ecology and Evolution. biomanipulation: Round Lake, Minnesota, the first Heidelberg: Springer-Verlag, pp. 60–81. two years. Freshwater Biology 14, 371–84. Werner, E.E. and Gilliam, J.F. (1984) The ontogenetic Sterner, R.W. (1990) The ratio of nitrogen to phosphorus niche and species interactions in size-structured popu- resupplied by herbivores: zooplankton and the algal lations. Annual Review of Ecology and Systematics competitive arena. American Naturalist 136, 209–29. 15, 393–26. Sterner, R.W. and Hessen, D.O. (1994) Algal nutrient Werner, E.E. and Hall, D.J. (1977) Competition and habi- limitation and the nutrition of aquatic herbivores. tat shifts in two sunfishes (Centrarchidae). Ecology 58, Annual Review of Ecology and Systematics 25, 1–29. 869–76. Stockner, J.G. and MacIsaac, E.A. (1996) British Werner, E.E. and Hall, D.J. (1979) Foraging efficiency and Columbia lake enrichment programme: two decades habitat switching in competing sunfishes. Ecology 60, of habitat enhancement for sockeye salmon. 256–64. Regulated Rivers 12, 547–61. Werner, E.E., Mittelbach, G.G., Hall, D.J. and Gilliam, Svärdson, G. (1949) Competition between trout and char J.F. (1983a) Experimental tests of optimal habitat use in (Salmo trutta and S. alpinus). Report from the Institute fish: the role of relative habitat profitability. Ecology of Freshwater Research, Drottningholm 29, 108–11. 64, 1525–39. Svärdson, G. (1976) Interspecific population dominance Werner, E.E., Gilliam, J.F., Hall, D.J. and Mittelbach, in fish communities of Scandinavian lakes. Report G.G. (1983b) An experimental test of the effects of from the Institute of Freshwater Research, predation risk on habitat use in fish. Ecology 64, Drottningholm 55, 144–71. 1540–8. Tonn, W.M., Magnuson, J.J., Rask, M. and Toivonen, Wetzel, R.G. (1979) The role of the littoral zone and detri- J. (1990) Intercontinental comparison of small tus in lake metabolism. Archiwe für Hydrobiologie forest lakes: fish assemblage patterns in Finland and Beihefte Ergebnisse der Limnologie 13, 145–61. Wisconsin. American Naturalist 136, 345–75. Winemiller, K.O. (1992) Ecomorphological diversifica- Tonn, W.M., Vandebos, R.E. and Paszkowski, C.A. (1995) tion in lowland freshwater fish assemblages from five Habitat on a broad scale: relative importance of immi- biotic regions. Ecological Monographs 61, 343–65. gration and extinction for small fish assemblages. Winemiller, K.O. and Leslie, M.A. (1992) Fish assem- Bulletin Francais de la Peche et de la Pisciculture blages across a complex tropical freshwater/marine 337/338/339, 47–61. ecotone. Environmental Biology of Fishes 34, 29–50. Townsend, C.R., Sutherland, W.J. and Perrow, M.R. Winemiller, K.O. and Ponwith, B.J. (1998) Comparative (1990) A modelling investigation of population cycles ecology of eleotrid fishes in Central American coastal in the fish Rutilus rutilus. Journal of Animal Ecology streams. Environmental Biology of Fishes 53, 373–84. 59, 469–85. Winfield, I.J. (1986) The influence of simulated aquatic Turner, A.M. and Mittelbach, G.G. (1990) Predator- macrophytes on the zooplankton consumption rate avoidance and community structure: interactions of juvenile roach, Rutilus rutilus, rudd, Scardinius among piscivores, planktivores, and plankton. Ecology erythophthalmus, and perch, Perca fluviatilis. Journal 71, 2241–54. of Fish Biology 29, 37–48. Vanni, M.J. (1996) Nutrient transport and recycling by Zaret, T.M. and Rand, A.S. (1971) Competition in tropi- consumers in lake food webs: implications for algal cal stream fishes: support for the competitive exclu- communities. In: G.A. Polis and K.O. Winemiller (eds) sion principle. Ecology 52, 336–42. Food Webs: Integration of Patterns and Dynamics. New York: Chapman & Hall, pp. 81–95. 16 Comparative Ecology of Marine Fish Communities

K. MARTHA M. JONES, DEAN G. FITZGERALD AND PETER F. SALE

16.1 INTRODUCTION First, fisheries management has required ecologi- cal information on the populations being fished Fish of a particular species do not occur in isolation and, to a considerable degree, has obtained that from others but as members of ecological commu- information by sampling the catch of the fishery nities. They interact with fish of other species, being managed (Evans and Grainger, Chapter 5, with invertebrate prey, with bird and mammal Volume 2; Sparre and Hart, Chapter 13, Volume 2). predators, and with the plants and sessile animals Second, study of the ecology of many fishes that that provide much of the physical structure of are of recreational or only marginal fishery interest their environments. It has been traditional to refer has necessarily used fishery-independent sam- to fish communities, meaning all fish species pling methods, but has still relied largely on in a location, and we follow this tradition while samples of fish removed from their environment noting that many of the important biotic interac- as the source of ecological data. Third, in certain tions are with non-fish organisms (cf. Pauly and environments where the approach is feasible, the Christensen, Chapter 10, Volume 2 for definition ecology of fishes has been explored by putting of an ecosystem). Our goal is to describe marine the scientist into the water. Methods similar fish communities, particularly to contrast com- to those used in terrestrial ecology are used to munities of fishes in temperate and tropical observe the fish in its habitat or to perform field demersal locations and in the pelagic realm. Are experiments. Each of these three approaches has structure and dynamics of fish communities in its own biases as well as its own methods. Each these different types of location similar or not focuses on its own set of questions and tells its own and why might differences exist? In contrast to version of reality (see Hart and Reynolds, Chapter Polunin and Pinnegar (Chapter 14, this volume) 1, this volume). we focus on the disparate perspectives of tropical Because these general methods are differen- and temperate communities given by different tially used among different environments, our sampling methods and on the role larval ecology pictures of fish communities are incomplete in plays in determining community structure. different ways in different places. Fisheries science Polunin and Pinnegar are more concerned with began with the need to manage commercial fish- how to establish trophic links and in understand- eries (Smith, Chapter 4, Volume 2). Its early devel- ing how ecosystem structure determines stability. opment was largely a north-temperate marine Our knowledge of the ecology of marine fishes endeavour, perhaps even a North Atlantic activity. has derived from three rather different approaches The study in temperate regions of the ecology of driven by different questions, and from different freshwater species and coastal marine species of ancestries that have yielded different perspectives. limited commercial value primarily used methods 342 Chapter 16 that kept scientists more or less dry and ensured 1984; Helfman et al. 1997; Holmlund and Hammer fish were more or less dead before they were exam- 1999). The bulk of this marine diversity (78%) ined. The use of techniques that put the scientist is found in coastal waters; 89% of these littoral– into the water has been most extensive in the trop- continental shelf species occur in tropical waters ics and the warm temperate zone. Our thesis will (Moyle and Cech 1996; Helfman et al. 1997). be that the differences in community structure Epipelagic fishes occupy the open ocean from the and dynamics reported among temperate, tropical surface down to 200m and constitute just 2.2% of and pelagic fish communities may be as much due the total number of marine species, although they to the different approaches used as to real differ- are the major component of the world’s marine fish ences in ecology. This chapter examines patterns landings. Coastal pelagics, particularly clupeoids, in biodiversity, habitat associations and produc- make up one-third of the total 80–90 million tion and recruitment dynamics, all in the context tonnes landed per year (Cushing 1995; Holmlund of community structure and dynamics, and at- and Hammer 1999; Pauly et al. 2000), and offshore tempts to identify real differences among these pelagics such as tunas and billfishes make up different environments. an additional 15% (Blaxter and Hunter 1982; Groombridge 1992). Deep pelagic fishes are those that live in the water column below 200m 16.2 BIODIVERSITY and constitute 8.6% of species. Deep demersal 16.2.1 Biogeographic divisions of species live on the bottom below 200m and the world’s oceans comprise 11% of total species. The number of species in these little explored regions is increas- The world’s oceans are expansive and continuous ing with time (Mooi and Gill, Chapter 3, this bodies of water, but compartmentalization into volume). somewhat distinct regions occurs on many spa- Because the focus of this chapter is to compare tial scales. On a large spatial scale, significant temperate and tropical fish communities, it is boundaries occur primarily due to physical oceano- inappropriate to attempt a survey of all biogeo- graphic factors, such as currents, fronts and other graphic regions. Instead, we focus on (i) tropical boundaries between water masses, the positions of demersal, chiefly coral-reef, communities of lit- the continents and suboceanic topography (Briggs toral fishes in the Caribbean and the vast Indo- 1974; Longhurst 1981; Sharp 1987; Helfman et al. West Pacific, (ii) temperate demersal communities 1997; Polunin and Pinnegar, Chapter 14, this of littoral species of the north temperate, and (iii) volume). Briggs (1974), the classic reference to epipelagic communities of temperate and sub- marine biogeography, recognizes cold temperate, tropical seas. warm temperate and tropical groupings of lit- toral–continental shelf species, as well as epi- pelagic, deep pelagic and deep demersal or abyssal 16.2.2 Levels of biodiversity in fish species. Temperate biogeographic regions are different systems restricted to single ocean basins or even to single Trend of diversity with latitude shorelines. Four tropical biogeographic regions are recognized for littoral fishes: Indo-West Pacific, Within large geographic regions, several patterns East Pacific, Western Atlantic–Caribbean and of faunal diversity have been reported in the lit- Eastern Atlantic–Mediterranean (see Gill and erature, with various explanations to account for Mooi, Chapter 2, this volume, who give a much these patterns (Paine 1966; Pianka 1966; Briggs broader categorization of biogeographic regions). 1974; Stevens 1989; Thresher 1991; Krebs 1994; The world contains in excess of 25000 de- Rosenzweig 1995). The most widely known and scribed species of fish (Eschmeyer et al. 1998), and accepted pattern of global faunal diversity lies in about 60% of the world’s fishes are marine (Nelson the trend of decreasing species diversity with in- Ecology of Marine Fish Communities 343

2500 WRI Briggs (1974) WCMC 2000 Fishbase Floeter and Gasparini (2000) ASIH 1500 y = –8.23x + 615.11 r2 = 0.15 P < 0.001 n 1000 = 261 Species richness

500

0 0 20406080 Latitude

Fig. 16.1 Plot of species richness against latitude for shore fish assemblages. Although the regression of declining richness with distance from the equator is significant, it accounts for only 15% of variance in the data. Data were compiled from various sources as shown: Briggs (1974); Floeter and Gasparini (2000); FISHBASE, ICLARM’s Fishbase website (http://www.cgiar.org/iclarm/fishbase/); WCMC, World Conservation Monitoring Centre website (http://www.wcmc.org.uk/marine/data/statistics/); WRI, World Resource Institute website (http://www.wri.org/wri/biodiv/marihome.html); ASIH, personal communications from colleagues at American Society of Icthyologists and Herpetologists conference, La Paz, Mexico, 2000, details available on request.

creasing latitude (Fischer 1960; Briggs 1974; Craig latitude, the trend in richness with latitude is seen 1984; Crossland 1988; Stevens 1989). This is not to be quite weak (Fig. 16.1). More comparative such a simple case in the marine environment. For studies on fishes are needed but, at present, it ap- example, marine mammals reach their maximal pears unlikely that a simple pattern with respect to diversity in polar regions, seaweeds reach their latitude will emerge. maximal diversity in temperate regions and corals reach their maximal diversity in the tropics (Gee Global clines, biodiversity hotspots and Warwick 1996). For most groups of marine organisms, the latitudinal pattern is unknown, Other aspects of global faunal diversity, including partly because the comparative data are lacking centres of distribution (‘hotspots’), patterns of en- (Gee and Warwick 1996). Several comparative demism and geographic patterns of composition of studies (Horn and Allen 1978; Stevens 1989; taxa, have attracted the attention of researchers in- Floeter and Gasparini 2000) have found evidence of terested in the diversity of marine fishes (Ekman the expected pattern of decreasing fish diversity 1953; Briggs 1970, 1974; Gilbert 1972; Springer with increasing latitude, excluding fishes that 1982; Thresher 1991). Here we review patterns in have extensive migrations. However, this confor- littoral fish faunas. mity with the general trend is due almost entirely to the much greater richness of fish communities Tropical littoral Because of the presence of coral in some tropical coastal, especially coral-reef, reefs, tropical littoral regions include locations locations. Indeed, when species lists from a broad possessing the highest faunal diversity, greatest geographic range of locations are plotted against gross photosynthesis and highest standing stock 344 Chapter 16 biomass of all marine ecosystems (Crossland et al. flect the considerable interregional barriers that 1991; Birkeland 1997). Yet coral reefs are uncom- have existed since the late Eocene. The relative mon, covering anywhere from 255000km2 numbers of species in each of the families varies (Spalding and Grenfell 1997) to 617000km2 (Smith substantially among regions so that some families 1978) depending on how narrowly they are defined. are of quite different importance in terms of This represents just 0.1–0.2% of ocean area scat- species richness even if represented in different tered through suitable coastal locales throughout regions (Thresher 1991); there are also measurable the tropics. The four littoral biogeographic regions differences of this type in taxonomic composi- in the tropics all contain coral reefs, and are sepa- tion among locations within the Indo-West Pacific rated chiefly by faunal characteristics of reef- (Springer 1982; Thresher 1991). There are also high associated fauna. Reef development is far more levels of endemism, and only a handful of cir- extensive in the Indo-West Pacific than elsewhere cumglobal coral-reef fishes (Briggs 1974; Thresher and very sparse in the Eastern Atlantic. Of the four, 1991; Allen and Robertson 1994; Robertson and the Indo-West Pacific and the Caribbean–Western Allen 1996). Recently, Jones et al. (2002) have Atlantic regions contain the greatest biodiversity shown that endemism is well distributed through- (Briggs 1974; Springer 1982; Thresher 1991; Floeter out the Indo-West Pacific and the Caribbean, and Gasparini 2000), although recent work is though with a tendency for it to be more pro- adding to the richness of the East Pacific region – nounced in the more isolated island groups within the coastal waters of Mexico and Central America the Pacific. and the few island groups east of the ‘East Pacific Barrier’ of deep oceanic waters (Allen and Temperate littoral In the north temperate region, Robertson 1994; Robertson and Allen 1996). boundaries are hard to fix because the change in Total fish species richness reaches in excess of fauna from one location to the next is gradual and 600 species in Caribbean locations and can exceed the ranges of many fishes fluctuate through- 2000 species in central Indo-West Pacific loca- out the year in response to seasonal climatic tions. In the latter, one can often see in excess of changes. Despite this, marine biogeographers have 100 species in a single dive. As one travels away recognized four main north temperate regions: from the centre of either region, in any direction, Mediterranean–Atlantic, North American species richness on coral reefs declines until it is Atlantic, North American Pacific and Asian Pacific comparable to species richness on warm temper- (Briggs 1974). The southern portions of these ate rocky shores, around 200 species or fewer regions possess representatives of tropical faunas, (Gilbert 1972; Sale 1980; Thresher 1991). This pat- such as Labridae, Serranidae, Haemulidae and tern of decline with distance from the centre of Lutjanidae; as one travels northward, fish more the region is both more pronounced and far better typical of cold water are found, such as Pleuronec- documented in the Indo-West Pacific, perhaps tidae, Gadidae and Cottidae. Species richness is simply because it is a much larger region than the notably lower than in tropical littoral regions and Caribbean–Western Atlantic. declines towards the north. In the North American There is a high degree of taxonomic similarity Atlantic region, for example, there are about 350 among the four regions at the level of family and littoral species south of Cape Hatteras, 200–250 even of genus; however they are well differentiated south of Cape Cod, about 225 in the Gulf of Maine, at the species level. This taxonomic similarity 100 in the Gulf of St Lawrence, 61 on the Labrador reflects the circumtropical oceanic system due coast and 34 in Greenland (Briggs 1974). Similar to the existence of the shallow Tethyan Seaway clines are evident in each of the other temperate during the time the tropical fish fauna was evolv- biogeographic regions. Briggs (1974) suggests that ing (Bellwood 1996, 1997). The differences among the North Pacific has been the primary centre of regions, particularly between the Caribbean– species evolution for north temperate regions and Western Atlantic and the Indo-West Pacific, re- has largely determined the faunal composition of Ecology of Marine Fish Communities 345 the less rich northern Atlantic. The Antarctic water techniques, such as underwater habitats or region has played a similar role in the south. submersibles, are very much more expensive. In- water methods also require good visibility, moder- ate current regimes and tolerable temperatures. Different faunas, different fisheries, While approaches that put scientists into the water different questions and approaches are used in cold temperate regions like Greenland Littoral fish faunas are far richer in the tropics than and the Gulf of Maine (Green 1975; Green et al. in the temperate zone. As a consequence, tropical 1984; Levin 1993), the physical conditions are not coastal fisheries involve localized multispecies conducive to this activity the way they are further harvests of the same suite of many species over south, and many parts of the continental shelf are broad areas (Munro 1996; Arreguin-Sanchez and outside the depth range for conventional SCUBA. It Manickchand-Heileman 1998; Pauly et al. 2000), is hardly surprising that the bulk of ecological while temperate fisheries are usually more closely research on fish in north temperate waters has used targeted to single species. In general, fisheries various types of gear to catch fish for subsequent research, which developed with commercial fish- analysis and that this research has focused on eries in the temperate zone (Smith, Chapter 4, questions of population ecology. Catch your fish, Volume 2), has focused on the interactions of a measure it, take scales, gut and gonad samples – this target species stock and the fishery, with limited could very well be the mantra driving ecological regard to other aspects of the environment or research on temperate fish species. the ecosystem of which the fished species is a By contrast, in the tropics it is unpleasant to be part. Yet the fundamental difference in the nature out on the water and not in it, and the highly of temperate and tropical coastal fisheries has diverse fish communities are predominantly in not led to different management approaches nor shallow clear waters, usually with moderate to different types of fisheries research, and tropi- current regimes. The high diversity of fish on coral cal coastal fisheries are poorly managed as a con- reefs as well as the ecological background of many sequence. The story is complicated further by primary researchers led initially to an interest in the fact that the far more limited financial, human the structure and dynamics of these assemblages and information resources available in develop- rather than to a study of the ecology of single ing countries diminish the capacity to do fisheries species (Sale 1988). Of course, in the same shallow research and management in the tropics (Sale tropical seas there are the turbid waters of man- 2002). groves and other estuaries in which direct observa- Scientists interested in fish communities rather tion is far more difficult. Methods conventionally than single fished species or fishery management used in temperate waters are employed there, have tended to study those communities most often to answer the same size, age, gut and gonad amenable to direct observation and experimenta- questions of population ecology. Yet we have yet to tion (but see Pauly and Christensen, Chapter 10, meet an ecologist who works on tropical coastal Volume 2). These are rarely the relatively depau- fishes and prefers working in mangroves and sea- perate groundfish communities of the North grass beds rather than coral reefs. Atlantic continental shelves where fisheries re- The community-level questions that dominated search developed. Instead, studies of marine fish early coral-reef studies were derived from terres- communities have predominantly examined the trial ecology rather than from study on fishes in diverse communities of shallow tropical and sub- other systems (Sale 1980). Even now, 50 years after tropical shores, and for some quite simple reasons serious coral-reef research on fishes began and 90 (Roberts 1996; Sale 1996). Use of SCUBA is effec- years since the first attempts by diving scientists tively limited to depths less than 25m for meaning- to make in-water observations of reef fishes, there ful ecological research, although use of Nitrox gas is remarkably little cross-referencing of tropical mixtures can extend that limit slightly. Other in- and temperate ecological research on fishes. 346 Chapter 16

16.2.3 Why and how are the ably because they are taxonomically predomi- tropics more diverse? nantly advanced perciforms (Leis 1991; Leis and Carson-Ewart 1997, 1998, 1999; Stobutzki 1998; Much of the earliest ecological research on coral- Stobutzki and Bellwood 1998). Reef fish commu- reef fishes was directed to the question of biodiver- nities, then, are metassemblages of species, each sity. Why were these systems so rich in species? protected from permanent local extinction by its What ecological processes permitted the con- metapopulation properties (Sale 1991b, 1996; tinued coexistence of such large numbers of see also Polunin and Pinnegar, Chapter 14, this species relative to temperate systems? Despite volume; Polunin, Chapter 14, Volume 2). This con- much effort (Sale 1980, 1991a), definitive answers sensus does not answer why these tropical faunas to these seductive questions do not yet exist. are so diverse but it diverts the question from the There is substantial evidence that reef fishes ecological to the evolutionary realm, leaving it as a can be highly specialized with respect to micro- problem for somebody else. Indeed, the question habitat (Sale 1980; Tolimieri 1995, 1998; Caselle may even be backwards. Should we not ask why and Warner 1996). That so many of them are quite it is that temperate fish communities are so sedentary once their pelagic larval phase has been depauperate? completed makes such specialization possible, and the topographically and biotically rich struc- ture of their environment enhances possibilities 16.3 HABITAT for microhabitat differentiation. A coral reef is ASSOCIATIONS structurally far more complex than any rocky reef, and both make the structural complexity of a soft 16.3.1 Demersal species sediment plain something to be pitied! However, the early expectation that such In demersal systems, our understanding of habi- microhabitat specialization would permit effec- tat is dominated by coral-reef investigations, tive habitat partitioning and stable coexistence of although there are some parallel investigations otherwise very similar species has not been borne in temperate systems (e.g. Holbrook et al. 1990; out (Sale 1991b). Indeed, there is growing evi- Tupper and Boutilier 1997). Demersal fishes select dence that coral-reef environments are subject to habitats for various reasons, including food re- frequent disturbances and are otherwise unpre- sources (Love and Ebeling 1978; Choat 1982; Jones dictable habitats (Connell 1978; Wellington and 1986; McCormick 1995), spawning and nesting Victor 1985; Williams 1986). There is also evi- sites (Fricke 1975; Walsh 1987; Danilowicz 1995; dence that in many instances reef fishes are not Warner 1995), sleeping sites and protection from limited by the availability of habitat or other predation (Shulman 1985; Jones 1988; Holbrook reef-based resources (Doherty 1983; Doherty and et al. 1990; Anderson 1994) and protection from Fowler 1994a,b; Caley et al. 1996; Hixon 1998). stressful environmental conditions. Some of the At present, if there is a consensus view, it is that patterns of fish–habitat associations may be con- reef fish exist as metapopulations composed of served in more depauperate systems, whereas numerous local open populations interconnected others involving the effect of species interactions by their larval stage. Reef fish larvae are nearly on these associations may not. uniformly pelagic for periods ranging from 1 week to several months but averaging 4 weeks, and are Coral reefs highly specialized for pelagic existence during this time. They also appear to be substantially more Relative to temperate demersal species, coral-reef capable pelagic organisms than the larvae of the species exist in a relatively stable environment temperate species that have been examined, prob- with respect to temperature, salinity and oxygen Ecology of Marine Fish Communities 347 conditions. Common disturbances they face are Fowler 1994a,b; Caley et al. 1996; Sale 1996). Habi- severe yet episodic, including hurricanes, point tat selection at the time of settlement can be pre- source siltation from localized erosion, or destruc- cise, and is facilitated in some cases by olfactory tive fishing methods (Bohnsack 1996). When com- cues deriving from adult fishes or coral species paring one reef locale to another within a region (Sweatman 1988; Sweatman and St John 1990; such as the Western Atlantic, for example, features Roberts 1996). However, two factors act to prevent that are relatively similar across reef habitats in- consistent microhabitat segregation by species: clude temperature, salinity, characteristically lim- (i) coral-reef habitats have high heterogeneity with iting nutrient levels, high water clarity and gross complex spatial arrangements of corals and resi- productivity, which is typically high. In contrast, dent fishes, thus providing multiple cues as attrac- features that may show marked differences are tants to various potential recruits; and (ii) recruits physiographic features such as live coral cover, are patchy in space and time and may therefore reef profile, extent of lagoon or shallow areas not be available to respond to the cues provided around the reef where mangroves and sea-grass (Caley et al. 1996). beds are found, and tidal range (Marshall 1985; Presently and historically, coral-reef managers Sharp 1987). have felt that limits of available habitat restrict Reef fish are strongly segregated by these physio- abundances of adult fish populations. This is likely graphic features at spatial scales of 10–100m (Sale incorrect much of the time. The recent paradigm 1991b), resulting in patterns of zonation where the shift in marine ecology that emphasizes the role of species found in zones such as forereef, reef flat and recruitment in setting local species abundances backreef are usually predictable and conserved, (Doherty and Fowler 1994a,b; Bohnsack and Ault although actual numbers within each species 1996; Caley et al. 1996; Sale 1996) may be begin- may fluctuate (Sale 1980). Another common pat- ning to filter into management of coral-reef fishes tern observed for coral-reef species is the increase in (Roberts 1996; Murray et al. 1999). Managers must the number of large individuals with increasing recognize the fact that presence of a pelagic larval depth. Fishing pressure usually attenuates with phase, and very high mortality during the pelagic depth and may amplify this pattern, but it is likely existence, ensure a large degree of serendipity with pre-existing since many unfished coral-reef species respect to the species and numbers of recruits that show the same trend. colonize adult habitats. A weak positive relationship often exists be- tween numbers of species present and habitat Other tropical demersal environments structure or structural complexity (Roberts 1996; Adjeroud et al. 1998; reviewed in Jones and Syms Other tropical demersal environments such as 1998). However, this is not a universal pattern mangrove, estuarine, sea-grass and ‘hardbottom’ even within coral-reef systems. Studies in differ- habitats have received more attention in recent ent tropical regions or with different assemblages years (Thayer et al. 1987; Bell and Pollard 1989; Ley of species have demonstrated strong positive et al. 1999; Lindeman and Snyder 1999). Observa- and negative effects of habitat structure as well tions of diverse fish assemblages in high densities as weak or no relationships (Sale 1991a; Jones in these environments, coupled with threats of and Syms 1998). anthropogenic stresses, have heightened concern At spatial scales of metres, microhabitat segre- and awareness of these highly productive habitats. gation only occurs among certain particularly Furthermore, there has been increasing evidence specialized species. Other processes, especially to suggest that mangrove, sea-grass and estuarine spatiotemporal variation in patterns of recruit- habitats serve as nursery grounds for important ment, may have a much greater effect on the fish sport and commercial fishery species (Livingston species we see at particular locations (Doherty and 1982; Heck and Thoman 1984; Sogard et al. 1989; 348 Chapter 16

Cabral and Costa 1999). Environmental factors problems in data interpretation (Pennington that influence fish distributional patterns in man- and Strømme 1998; Parrish 1999; Engås et al. 2000; groves and estuaries include gradients in salinity, Sparre and Hart, Chapter 13, Volume 2). Overall, water clarity, depth, abundance of prey and tem- despite the difficulty, direct observations by perature (Ley et al. 1999). SCUBA or submersible are invaluable if a detailed picture of habitat responses is to be gained. In contrast to coral-reef systems, temperate Temperate demersal environments demersal environments show marked spatial Most temperate studies of habitat associations differences in temperature, oxygen, salinity and have focused on the effect on species distribution primary production (Coutant 1990; Sharp 1987). In of a primary environmental factor, such as depth, fact, one of the steepest latitudinal temperature structural relief or substratum type. Recent find- gradients in the world, at 1°C for every 1° latitude, ings are revealing many relationships between occurs along the east coast of North America, com- temperate fishes and their environment, and con- bining in some parts with an annual range in sea tinental shelves are no longer thought of as simple surface temperature of up to 20°C (Conover 1992). homogeneous environments with uniform assem- Such pronounced latitudinal and seasonal changes blages of fishes (Craig 1984; Grimes et al. 1986; have led to numerous studies on the relationship of Hobson 1986; Pihl and Ulmestrand 1993; Gregory temperature to commercially important species, and Anderson 1997; Rhodes 1998; Swain et al. for example for groundfish (Swain et al. 1998) 1998). For example, beam trawl surveys in the Atlantic cod, Gadus morhua, (Pihl and Umestrand northeast Atlantic by Rogers et al. (1999) found 1993) and Alaskan coastal fishes (Craig 1984). that more heterogeneous, rocky substrata sup- These studies have frequently identified broad- ported a more diverse fauna of smaller-sized fish, scale spatial structure in fish populations. Yet in such as gurnards (Trigla spp.), sole (Solea solea) and addition to temperature or other physicochemical elasmobranchs, than more homogeneous, muddy factors, there is often significant habitat structure bottoms. These results were similar to those of created by the sponges and other sessile inver- Platell and Potter (1999), who found partial segre- tebrates that grow even where the substratum is gation of Triglidae and Pempheridae in temperate uniformly sand or clay. Indeed, awareness that muddy and rocky bottoms. Tupper and Boutilier temperate continental shelves frequently contain (1997) have demonstrated the importance of significant structure of importance to fishes ap- habitat on various aspects of temperate reef fish pears to have coincided with the realization that ecology, including settlement, growth, predation commercial trawling in many regions is now so in- risk and survival. Ebeling and Hixon (1991) argued tense that most patches of substratum are trawled convincingly for the general similarity of patterns annually, with resulting serious degradation to in temperate rocky reef fish assemblages and those structure (Dayton et al. 1995; Watling and Norse on coral reefs. 1998; Turner et al. 1999; Kaiser and Jennings, Identifying detailed patterns of association of Chapter 16, Volume 2). fishes with habitat structure is difficult when For example, directed movements of Atlantic techniques like trawling are used to determine dis- cod related to temperature have been associated tribution of fish (Larkin 1996; Walters et al. 1997; with aggregations for feeding, overwintering and Pennington and Strømme 1998; Millar and Fryer spawning among inshore and offshore habitats 1999). Trawl efficiency varies with habitat and (Templeman 1966; Garrod 1988; Serchuk and with overall abundance of fish (Godø et al. 1999), Wigley 1992). While the earlier perspective was and discrepancies between path of vessel and one of broadly uniform distribution over entire off- paths of towed instruments introduce significant shore banks and embayments, more recent studies additional errors (Engås et al. 2000). Patchy dis- have identified meso- and small-scale genetic tributions of fish within habitats cause further heterogeneity that is being attributed to known Ecology of Marine Fish Communities 349 oceanic features besides cod biology and ecology, ity of species this range, or a portion of it, may be and implies far greater association of individuals defended as a territory. Many fish may undergo with specific places than previously suspected. By periodic foraging and spawning migrations away using microelemental analysis of otoliths, individ- from their home range, and in some species these ual spawning aggregations have been shown to movements may cover tens of kilometres. On be associated with unique elemental fingerprints; coral reefs, home ranges are dependent on fish size these infer population structure on a scale as small and species and the typical relationship is for larger as inshore and offshore zones of individual bays species to move over larger home ranges (Sale and within specific locations on Georges Bank, the 1978). Familiarity with an area that comes with Grand Banks and the Labrador Shelf (Ruzzante the establishment of home ranges offers many ad- et al. 1999, 2000; Campana et al. 2000). Stock- vantages, such as accurate knowledge of feeding, specific spawning occurs in the presence of mating and shelter sites (Braithwaite 1998; Kramer topographically induced eddies and gyre-like cir- and Chapman 1999). These advantages would not culation that result in the consistent deposition be expected to differ substantially between tropi- and retention of eggs and larvae in a particular cal and temperate systems (Barrett 1995), although habitat. the tendency for year-round residency may de- Habitats thought to be contiguous are now being crease with increasing latitude because some tem- described as patchy due to the presence of sub- perate regions experience large seasonal variation marine channels, saddles and trenches that iso- in water temperature, forcing reef residents into late populations. Such physical features can also deeper waters to avoid environmental extremes promote coordinated movements of large num- (Barrett 1995). The findings for cod presented bers of fish among habitats. These are termed earlier provide support for this hypothesis. Home ‘migration highways’. Habitat structure can also ranges may be very restricted, exemplified by focus spawning aggregations, termed ‘spawning gobies (Gobiidae) and damselfish (Pomacentridae) columns’, at specific sites (Rose 1993). Described that commonly reside within 1–2m2 around a in these contexts, Atlantic cod appear to resemble coral head, and by pricklebacks (Stichaeidae) and the larger snappers and groupers of coral reefs that sculpins (Cottidae) that occupy small temper- live closely associated with a topographically rich ate tidepools (Barrett 1995; Helfman et al. 1997; substratum but undertake substantial annual Kramer and Chapman 1999). Intermediate ranges migrations to traditional spawning locations tens of a few hundred square metres characterize of kilometres distant in characteristic outer reef many kelp-bed and reef species (see Kramer and slope locations (Colin 1992; Sadovy et al. 1994; Chapman 1999 for examples). Domeier and Colin 1997; Zeller 1997, 1998). Such work suggests that the well-recognized similar- 16.3.2 Habitats for pelagic ities between fish communities on coral reefs species: Water masses, fronts and rocky reefs (Ebeling and Hixon 1991) are likely and eddies extendable to many other temperate demersal communities. When managers and scientists discuss habitat with reference to pelagic fisheries, they generally refer to physical oceanography such as frontal Home ranges and territories zones, oxygen, nutrient, salinity and temperature Overall, we suggest there is a general trend for gradients (Sharp 1987; Cushing 1995). Pelagic re- long-term site attachment in demersal species gions are characterized by large volumes of water in all environments. Long-term site attachment that lack obvious physical structure, have high refers to the occupation of an area by an individual solar insolation and have variable production that for its entire adult life. In most this will take the can be very high in regions of upwelling or conver- form of a long-term home range, while in a minor- gence of major currents (Walsh 1981; Mann 1993; 350 Chapter 16

Cushing 1995; Verity and Smetacek 1996). Fronts and vascular plants. The role of detritivores that or upwelling areas represent concentrated pro- remobilize carbon into food webs is important but ductivity in temperate marine waters and can be also easily forgotten (Wood 1967; DeAngelis et al. seen from aircraft or satellites via infrared radio- 1989; Daly and Smith 1993). Marine fish commu- metry, provided there are temperature differences nities are supported by three different production (Cushing 1995). These frontal zones represent cycles that are differentially important in different boundaries between water masses that differ regions. in temperature, salinity or both, and possess an The water column supports two alternative accumulation of material including planktonic phytoplankton production cycles. Both are ex- organisms and fishes that feed on the plankton ploited by zooplankton and fishes (Daly and Smith (Walsh 1981; Sherman 1994). Although frontal 1993; Cushing 1995; Mann and Lazier 1996). First, zones change rapidly (Grioche and Koubbi 1997), in regions where upwelling returns nutrients to they are predictable through the use of oceano- surface waters, bursts of primary production of graphic modelling (Sharp 1987; Cushing 1995). phytoplankton, mainly diatoms, are promoted. Up- There is abundant evidence that epipelagic fishes welling can be the product of wind stress from aggregate at these discontinuities and that target- above or tidal currents impinging on the bottom ing such zones is an effective fishery practice from below and may depend on submarine topog- (Kirby et al. 2000). Fronts also may occur between raphy. Such zones occur in association with the masses of water differing in salinity but not continental shelves, estuaries, tides and current temperature. In these cases the water masses may boundaries. The phytoplankton are kept in an en- be indistinguishable from satellite radiometry vironment rich in essential nutrients and light. data, yet they may differ significantly in fauna Maximal rates of primary production result. Phyto- (Walsh 1981; Sharp 1987; Castillo et al. 1996). plankton production can be used as a source of food directly by vertebrate grazers, such as herbivorous fishes, or indirectly by zooplankton that are then available for consumption by planktivorous fishes. 16.4 DIFFERENCES IN Many fishes have evolved specialized gill arches or TROPICAL AND TEMPERATE jaws that allow for efficient and maximal consump- PRODUCTION CYCLES tion of available phytoplankton or zooplankton resources. Any production not consumed by fish is Production cycles and trophic relationships are recycled back into the food web by the microbial covered by Pauly and Christensen (Chapter 10, loop or settles to the bottom and is consumed by the Volume 2) and Polunin and Pinnegar (Chapter 14, benthic invertebrate community (Daly and Smith this volume) respectively, so we limit comment to 1993; Cushing 1995; Mann and Lazier 1996). The geographic trends in fish production. Fish commu- majority of offshore fisheries in temperate and nity structure and population dynamics are linked tropical systems are sustained by this production to physical oceanography by primary produc- cycle. tion cycles (Woods 1988; Cushing 1995; Mann and The second production cycle is associated with Lazier 1996). The primary production within each zones that experience well-stratified conditions system shapes the characteristic annual produc- through extended periods, such as tropical oceanic tion efficiency and quantity of fishery stocks waters year-round and most temperate waters and the communities they comprise (Woods 1988; during the summer months. In these zones, lack Daly and Smith 1993; Mann 1993; Cushing 1995; of upwelling allows the euphotic zone to become Mann and Lazier 1996). The primary source of depleted of nutrients. Primary production is at energy is incoming solar radiation and carbon fixa- relatively low rates and dominated by bacteria and tion is performed by phytoplankton, macroalgae small flagellates. These producers are predomi- Ecology of Marine Fish Communities 351 nantly taxa that are small in size. They tend to sup- that variation in the recruitment of new juvenile port several trophic levels of zooplankton and are fishes plays a significant role in determining the relatively quickly recycled by the microbial loop. structure of fish communities. Myers (Chapter 6, They are considered to be a poor source of energy this volume) provides a comprehensive review of for the production of fishes. For weakly stratified recruitment and density-dependence in fish popu- water, production of phytoplankton, again domi- lations. We use the term ‘recruitment’ not in the nated by diatoms, will be favoured over production fisheries but in the ecological sense of ‘addition of bacteria and flagellates when conditions pro- of a new cohort to a population’ (Caley et al. 1996). mote a low ratio of respiration to photosynthesis It is an important process in communities such as (Daly and Smith 1993; Cushing 1995; Mann and those on coral reefs, which are made up of species Lazier 1996). with pelagic larval stages and at least a modest In addition to these two water-column produc- degree of mixing of larvae from adjacent breeding tion cycles, demersal fishes in shallow waters, populations. In such systems, the process of typically <30m, have access to benthic primary ‘settlement’ from the pelagic stage, and the usually production in the form of algae and sea-grasses. synchronous metamorphosis from a larval to a Demersal primary production can operate at the juvenile phenotype, are dramatic moments in highest known rates in very shallow systems with life history when one lifestyle is exchanged for efficient nutrient recycling mechanisms, and is of another radically different one (Leis 1991). The major importance as a food source for herbivores in great majority of other marine fishes also have coral-reef communities (Hatcher 1990). Herbivo- dispersive larval stages that differ dramatically rous fishes that feed on demersal algae and sea- from the adults, so these species may also ex- grasses typically comprise 25–40% of coral-reef perience pronounced variations in recruitment to fish biomass (Polunin 1996). This is in marked the juvenile stage. contrast to their near absence on cold temperate reefs, where demersal production enters the fish 16.5.1 Spatiotemporal variation in community principally by way of invertebrate recruitment to reef habitats intermediates. The cold temperate, Southern Hemisphere Odacidae are particularly capable A number of studies in various coral reef regions digesters of algal material and an exception to have now demonstrated a conspicuous spatiotem- the rule that algal herbivory by fish is a tropical poral variation in recruitment of fish to the juve- lifestyle (Choat and Clements 1998). As a conse- nile stage (Williams and Sale 1981; Doherty and quence, tropical demersal habitats support fish Fowler 1994a,b; Caselle and Warner 1996; Tolim- communities that depend on benthic as well as ieri et al. 1998; reviewed in Doherty and Williams pelagic production to a degree not seen in the 1988; Caley et al. 1996). The variation can fre- temperate zone. The high rates of production quently be orders of magnitude in extent, both in of fish on coral reefs and other tropical shore space and time. Little of the variation among years waters would not be possible without this direct and locations is accounted for by variation in habi- use of demersal primary production. tat quality or resident population size. Doherty and Fowler (1994a,b) showed for one long-lived pomacentrid fish, Pomacentrus moluccensis, that 16.5 VARIATION IN differing patterns of recruitment success over a RECRUITMENT DYNAMICS decade left measurable and different ‘recruitment signatures’ in the age structure of populations at We include brief discussion of recruitment each of seven nearby coral reefs. With the excep- dynamics in this chapter because there is a real tion of one reef, which received particularly strong possibility, at least in tropical demersal systems, rates of recruitment and on which the population 352 Chapter 16

‘might’ have demonstrated some compensatory recruitment of reef fishes to the demersal popula- density-dependent mortality, the data strongly tion. Nevertheless, given that fishery recruitment supported an absence of any compensatory mortal- is so frequently highly variable in space and time, ity to adjust the varying input of new fishes. These it is reasonable to assume that entrance into earlier populations were ‘recruitment limited’ (Doherty life stages of fishery species will also be highly vari- 1983; Victor 1983, 1986) in the sense that their able, as on coral reefs, and there is some direct evi- abundances and age compositions were set by the dence that this supposition is correct. varying input of new settlers. While the lower diversity and stronger domi- It follows, from this and other studies of recruit- nance of most temperate communities will ame- ment variation, that community structure may be liorate the community-level effect, it follows that continuously modified by the varying recruitment at local scales the communities of fishes that occu- of many of the component species. This is clearly py demersal habitat on temperate shores may also the case for the small but isolated communities of be expected to have their composition constantly fish that occur in habitat composed of small patch modified, in relative abundances if not in species reefs (Sale 1991b; Sale et al. 1994). However, it is composition, due to the varying pattern of delivery very likely also the case for the larger communities of new cohorts of their component species. The ex- of fish on more contiguous reef habitat. The dif- tent of this variation is not well known, partly be- ficulty of accurately censusing communities over cause it has seldom been looked for. Its potential large areas of contiguous reef and the likelihood importance lies in the impact on such aspects of that spatial averaging will reduce the extent of community dynamics as the trophic web that variability seen in large study sites make this diffi- transfers nutrients and energy through the com- cult to test in the way it has been tested on small munity. We suggest that it is time to recognize that patch reefs (but see Ault and Johnson 1998). such ecosystem properties as patterns of transfer Accepting the notion that community com- of energy and nutrients must be seen as flexible position is continuously modified by recruitment and constantly changing. Predator–prey dynamics variation requires acceptance of the idea that reef may also change as particular prey or predator fish populations are normally not at limits set by species change in abundance relative to those they the resources available in reef habitat, although interact with. those limits are undoubtedly reached for some species at some times and places. Indeed, reef fish ecologists are currently strongly divided over the 16.6 CONCLUSIONS whole question of population regulation, and some are not convinced that density-dependence is a In this chapter we have used the richer body of particularly important regulatory phenomenon in-water direct data from tropical demersal sys- in such systems (Sale 1991b; Hixon and Carr 1997; tems to infer patterns and processes likely to be Chesson 1998; Hixon 1998; Sale and Tolimieri present in temperate demersal and pelagic fish 2000). communities. We have used direct data from these other systems to confirm or modify these infer- 16.5.2 Temporal variation in ences. Our sense is that there are substantial recruitment to temperate populations similarities among demersal fish communities and that a greater cross-referencing of work on The substantial variation in recruitment has been demersal systems should take place than has been the bane of fisheries management from the outset the case in the past. We suspect that the pelagic (Hjort 1914; Myers, Chapter 6, this volume). Here, habitat makes for somewhat different community recruitment is used in the fisheries sense, as ‘en- organization. trance of new individuals into the fishable popula- Biotic diversity of fishes does not follow any tion’, and occurs substantially later in life than the simple latitudinal cline, despite the existence of a Ecology of Marine Fish Communities 353 significant negative trend with latitude in most herbivorous taxa and the temperate demersal and sets of data on demersal shorefish. Variation in di- pelagic faunas that depend on intermediary inver- versity of shorefishes along the equator is just as tebrate herbivores. great, and some isolated islands close to the equa- tor have fish faunas no richer than north temperate coasts. The order of magnitude variation in species REFERENCES richness among shorefish faunas is likely to be a re- sult of history rather than a direct response to envi- Adjeroud, M., Letourneur, Y., Porcher, M. and Salvat, B. ronmental conditions. (1998) Factors influencing spatial distribution of Pelagic environments are structured by tem- fish communities on a fringing reef at Mauritius, S.W. perature, salinity and currents, at both small and Indian Ocean. Environmental Biology of Fishes 53, 169–82. large scales, although the structures are transient. Allen, G.R. and Robertson, D.R. (1994) Fishes of the Individual fishes, their populations and com- Tropical Eastern Pacific. Honolulu: University of munities respond to this structure and restrict Hawaii Press. their activities appropriately. We are beginning Anderson, T.W. (1994) Role of macroalgal structure in the to understand the complexity of these responses. distribution and abundance of a temperate reef fish. However, while knowledge of the details of these Marine Ecology Progress Series 3, 279–96. Arreguin-Sanchez, A. and Manickchand-Heileman, S. responses may enable a more effective targeting of (1998) The trophic role of lutjanid fish and impacts fishery species, it is unlikely that management can of their fisheries in two ecosystems in the Gulf of be implemented at other than broader scales as Mexico. Journal of Fish Biology 53 (Suppl. A), 143–53. done at present. Ault, T.R. and Johnson, C.R. (1998) Spatially and tempo- There is good evidence that demersal fish re- rally predictable fish communities on coral reefs. spond predictably to structural habitat features Ecological Monographs 68, 25–50. and reside within appropriate habitats. We know Barrett, N.S. (1995) Short- and long-term movement pat- terns of six temperate reef fishes (Families Labridae far more about this phenomenon in tropical coral- and Monacanthidae). Marine and Freshwater Re- reef environments than in temperate or pelagic search 46, 853–60. systems. The microhabitat structure created by Bell, J.D. and Pollard, D.A. (1989) Ecology of fish assem- sponges and other sessile invertebrates on other- blages and fisheries associated with seagrasses. In: wise featureless sandy or silty plains is only now A.W.D. Larkum, A.J. McCoomb and S. Shepherd (eds) becoming recognized as important for temperate Biology of Seagrasses: A Treatise on the Biology of demersal fish. In pelagic systems, fish respond to Seagrasses with Special Reference to the Australian Region. Amsterdam: Elsevier, pp. 565–609. the large- and smaller-scale structure of the water Bellwood, D.R. (1996) The Eocene fishes of Monte Bolca: mass due to physical oceanography. Our knowl- the earliest coral reef fish assemblage. Coral Reefs 15, edge of these patterns has grown with our ability to 11–19. define the physical structures using synoptic sam- Bellwood, D.R. (1997) Reef fish biogeography: habitat as- pling of fish and physical structure and via remote sociations, fossils and phylogenies. Proceedings of the sensing. We suggest that the interactions of demer- 8th International Coral Reef Symposium 1, 379–84. Birkeland, C. (1997) Symbiosis, fisheries and economic sal marine fishes and their habitats are defined development on coral reefs. Trends in Ecology and by spatial homogeneity of habitats, seasonality of Evolution 12, 364–7. conditions and the spatial scales at which fish live Blaxter, J.H.S. and Hunter, J.R. (1982) The biology of their lives. clupeoid fishes. Advances in Marine Biology20, 1–223. The highly variable pattern of recruitment from Bohnsack, J.A. (1996) Maintenance and recovery of reef larval to juvenile stages is apparent across ma- fishery productivity. In: N.V.C. Polunin and C.M. rine fish species. Its consequences for community Roberts (eds) Reef Fisheries. London: Chapman & Hall, pp. 1–14. structure will depend on the interaction with habi- Bohnsack, J.A. and Ault, J.S. (1996) Management strate- tat complexity. Productivity of fish is notably dif- gies to conserve marine biodiversity. Oceanography 9, ferent among tropical demersal faunas with their 73–82. 354 Chapter 16

Braithwaite, V.A. (1998) Spatial memory, landmark use Crossland, C.J. (1988) Latitudinal comparisons of coral and orientation in fish. In: S. Healy (ed.) Spatial Repre- reef structure and function. Proceedings of the 6th In- sentation in Animals. New York: Oxford University ternational Coral Reef Symposium 1, 221–6. Press, pp. 86–102. Crossland, C.J., Hatcher, B.G. and Smith, S.V. (1991) Role Briggs, J.C. (1970) Tropical shelf zoogeography. Proceed- of coral reefs in global ocean production. Coral Reefs ings of the California Academy of Sciences Fourth 10, 55–64. Series 38, 131–8. Cushing, D.H. (1995) Population Production and Regu- Briggs, J.C. (1974) Marine Zoogeography. New York: lation in the Sea: A Fisheries Perspective. Cambridge: McGraw-Hill. Cambridge University Press. Cabral, H. and Costa, M.J. (1999) Differential use of nur- Daly, K.L. and Smith, W.O. Jr (1993) Physical–biological sery areas within the Tagus estuary by sympatric soles, interactions influencing marine plankton production. Solea solea and Solea senegalensis. Environmental Annual Review of Ecology and Systematics 24, Biology of Fishes 56, 389–97. 555–85. Caley, M.J. Carr, M.H., Hixon, M.A., Hughes, T.P., Jones, Danilowicz, B.S. (1995) Spatial patterns of spawning in G.P. and Menge, B.A. (1996) Recruitment and the local the coral reef fish, Dascyllus albisella. Marine Biology dynamics of open marine populations. Annual Review 122, 145–55. of Ecology and Systematics 27, 477–500. Dayton, P.K., Thrush, S.F., Agardy, M.T. and Hofman, Campana, S.E., Chouinard, G.A., Hanson, J.M., Fréchet, R.J. (1995) Environmental effects of marine fishing. A. and Brattey, J. (2000) Otolith elemental fingerprints Aquatic Conservation Marine and Freshwater Eco- as biological tracers of fish stocks. Fisheries Research systems 5, 205–32. 46, 343–57. DeAngelis, D.L., Mulholland, P.J., Palumbo, A.V., Caselle, J.E. and Warner, R.R. (1996) Variability in re- Steinman, A.D., Huston, M.A. and Elwood, J.A. (1989) cruitment of coral reef fishes: the importance of habi- Nutrient dynamics and food-web stability. Annual tat at two spatial scales. Ecology 77, 2488–504. Review of Ecology and Systematics 20, 71–95. Castillo, J., Barbieri, M.A. and Gonzalez, A. (1996) Rela- Doherty, P.J. (1983) Tropical territorial damselfishes: is tionships between sea surface temperature, salinity, density limited by aggression or recruitment? Ecology and pelagic fish distribution off northern Chile. ICES 64, 176–90. Journal of Marine Science 53, 139–46. Doherty, P.J. and Fowler, A.J. (1994a) An empirical test Chesson, P.L. (1998) Recruitment limitation: a theoreti- of recruitment limitation in a coral reef fish. Science cal perspective. Australian Journal of Ecology 23, 263, 935–9. 234–40. Doherty, P.J. and Fowler, A.J. (1994b) Demographic Choat, J.H. (1982) Fish feeding and the structure of consequences of variable recruitment to coral reef benthic communities in temperate waters. Annual fish populations: a congeneric comparison of two Review of Ecology and Systematics 13, 423–49. damselfishes. Bulletin of Marine Science 54, 297–313. Choat, J.H. and Clements, K.D. (1998) Vertebrate herbi- Doherty, P.J. and Williams, D.McB. (1988) The replenish- vores in marine and terrestrial environments: a nutri- ment of coral reef fish populations. Oceanography and tional ecology perspective. Annual Review of Ecology Marine Biology Annual Review 26, 487–551. and Systematics 29, 375–403. Domeier, M.L. and Colin, P.L. (1997) Tropical reef fish Colin, P.L. (1992) Reproduction of the Nassau grouper, spawning aggregations: defined and reviewed. Bulletin Epinephelus striatus (Pisces: Serranidae) and its rela- of Marine Science 60, 698–726. tionship to environmental conditions. Environmental Ebeling, A.W. and Hixon, M.A. (1991) Tropical and Biology of Fishes 34, 357–77. temperate reef fishes: comparison of community Connell, J.H. (1978) Diversity in tropical rain forests and structures. In: P.F. Sale (ed.) The Ecology of Fishes on coral reefs. Science 199, 1302–10. Coral Reefs. San Diego, CA: Academic Press, pp. Conover, D.O. (1992) Seasonality and the scheduling of 509–63. life history at different latitudes. Journal of Fish Biol- Ekman, S. (1953) Zoogeography of the Sea. London: ogy 41 (Suppl. B), 161–78. Sidgwick and Jackson. Coutant, C.C. (1990) Temperature–oxygen habitat for Engås, A., Godø, O.R. and Jørgensen, T. (2000) A compari- freshwater and coastal striped bass in a changing cli- son between vessel and trawl tracks as observed by the mate. Transactions of the American Fisheries Society ITI trawl instrumentation. Fisheries Research 45, 119, 240–53. 297–301. Craig, P.C. (1984) Fish use of coastal waters of the Eschmeyer, W.N., Ferraris, C.J. Jr. and Hoang, M.D. Alaskan Beaufort Sea: a review. Transactions of the (1998) Catalog of Fishes. San Francisco: California American Fisheries Society 113, 265–82. Academy of Sciences. Ecology of Marine Fish Communities 355

Fischer, A.G. (1960) Latitudinal variations in organic Hixon, M.A. (1998) Population dynamics of coral- diversity. Evolution 14, 64–81. reef fishes: controversial concepts and hypotheses. Floeter, S.R. and Gasparini, J.L. (2000) The southwestern Australian Journal of Ecology 23, 192–201. Atlantic reef fish fauna: composition and zoogeogra- Hixon, M.A. and Carr, M.H. (1997) Synergistic predation, phic patterns. Journal of Fish Biology 56, 1099–114. density dependence, and population regulation in Fricke, H.W. (1975) Evolution of social systems through marine fish. Science 277, 946–9. site attachment in fish. Zeitschrift für Tierpsychologie Hjort, J. (1914) Fluctuations in the great fisheries of 39, 206–10. northern Europe viewed in the light of biological re- Garrod, D.J. (1988) North Atlantic cod: fisheries and search. Rapports et Procès-verbaux des Réunions du management to 1986. In: J.A. Gulland (ed.) Fish Popu- Conseil International pour l’Exploration de la Mer 20, lation Dynamics, 2nd edn. New York: John Wiley & 1–228. Sons, pp. 185–219. Hobson, E.S. (1986) Predation on the Pacific sand lance, Gee, J.M. and Warwick, R.M. (1996) A study of global bio- Ammodytes hexapterus (Pisces: Ammodytidae), dur- diversity patterns in the marine motile fauna of hard ing the transition between day and night in southeast- substrata. Journal of the Marine Biological Associa- ern Alaska. Copeia 1986, 223–6. tion of the UK 76, 177–84. Holbrook, S.J., Carr, M.H., Schmitt, R.J. and Coyer, J.A. Gilbert, C.R. (1972) Characteristics of the Western (1990) Effect of giant kelp on local abundance of reef Atlantic reef-fish fauna. Quarterly Journal of the fishes: the importance of ontogenetic resource require- Florida Academy of Sciences 35, 130–44. ments. Bulletin of Marine Science 47, 104–14. Godø, O.R., Walch, S.J. and Engås, A. (1999) Investigating Holmlund, C.M. and Hammer, M. (1999) Ecosystem density-dependent catchability in bottom-trawl sur- services generated by fish populations. Ecological veys. ICES Journal of Marine Science 55, 292–8. Economics 29, 253–68. Green, J.M. (1975) Restricted movements and homing Horn, M.H. and Allen, L.G. (1978) A distributional of the cunner, Tautogolabrus adspersus (Walbaum) analysis of California coastal marine fishes. Journal of (Pisces: Labridae). Canadian Journal of Zoology 49, Biogeography 5, 23–42. 1497–9. Jones, G.P. (1986) Food availability affects growth in a Green, J.M., Martel, G. and Martin, D.W. (1984) Com- coral reef fish. Oecologia 70, 136–9. parisons of the feeding activity and diets of male and Jones, G.P. (1988) Experimental evaluation of the effects female cunners, Tautogolabrus adspersus (Pisces: of habitat structure and competitive interactions on Labridae). Marine Biology 84, 7–11. the juveniles of two coral reef fishes. Journal of Experi- Gregory, R.S. and Anderson, J.T. (1997) Substrate selec- mental Marine Biology and Ecology 123, 115–26. tions and use of protective cover by juvenile Atlantic Jones, G.P. and Syms, C. (1998) Disturbance, habitat cod, Gadus morhua, in inshore waters off Newfound- structure, and the ecology of fishes on coral reefs. land. Marine Ecology Progress Series 146, 9–20. Australian Journal of Ecology 23, 287–97. Grimes, C.B., Able, K.W. and Jones, R.S. (1986) Tilefish, Jones, G.P., Caley, M.J. and Munday, P.L. (2002) Rarity in Lopholatilus chamaeleonticeps, habitat, behavior, coral reef fish communities. In: P.F. Sale (ed.) Coral and community structure in Mid-Atlantic and Reef Fishes. Dynamics and Diversity in a Complex southern New England waters. Environmental Ecosystem. San Diego, CA: Academic Press, pp. Biology of Fishes 15, 273–92. 81–101. Grioche, A. and Koubbi, P. (1997) A preliminary study of Kirby, D.S., Fiksen, Ø. and Hart, P.J.B. (2000) A dynamic the influence of a coastal frontal structure on ichthy- optimisation model for the behaviour of tunas at ocean oplankton assemblages in the English Channel. ICES fronts. Fisheries Oceanography 9, 328–42. Journal of Marine Science 54, 93–104. Kramer, D.L. and Chapman, M.R. (1999) Implications Groombridge, B. (ed.) (1992) Global Biodiversity: Status of fish home range size and relocation for marine of the Earth’s Living Resources. London: Chapman & reserve function. Environmental Biology of Fishes 55, Hall. 65–79. Hatcher, B.G. (1990) The primary productivity of coral Krebs, C.J. (1994) Ecology: The Experimental Analysis of reefs: a hierarchy of pattern and process. Trends in Distribution and Abundance, 4th edn. New York: Ecology and Evolution 5, 149–55. Harper Collins. Heck, K.L. and Thoman, T.A. (1984) The nursery role of Larkin, P.A. (1996) Concepts and issues in marine ecosys- seagrass meadows in the upper and lower reaches of the tem management. Reviews in Fish Biology and Fish- Chesapeake Bay. Estuaries 7, 70–92. eries 6, 139–64. Helfman, G.S., Collette, B.B. and Facey, D.E. (1997) The Leis, J.M. (1991) The pelagic phase of coral reef fishes: Diversity of Fishes. Boston: Blackwell Science. larval biology of coral reef fishes. In: P.F. Sale (ed.) The 356 Chapter 16

Ecology of Fishes on Coral Reefs. San Diego, CA: Acad- Munro, J.L. (1996) The scope of tropical fisheries and emic Press, pp. 183–230. their management. In: N.V.C. Polunin and C.M. Leis, J.M. and Carson-Ewart, B.M. (1997) In situ swim- Roberts (eds) Reef Fisheries. London: Chapman & Hall, ming speeds of the late stage larvae of some Indo- pp. 1–14. Pacific coral reef fishes. Marine Ecology Progress Murray, S.N. et al. (1999) No-take reserve networks: sus- Series 159, 165–74. taining fishery populations and marine ecosystems. Leis, J.M. and Carson-Ewart, B.M. (1998) Complex be- Fisheries 23(11), 11–25. haviour by coral-reef fish larvae in open-water and Nelson, J.S. (1984) Fishes of the World, 2nd edn. New near-reef pelagic environments. Environmental Biol- York: John Wiley & Sons. ogy of Fishes 53, 259–66. Paine, R.T. (1966) Food web complexity and species Leis, J.M. and Carson-Ewart, B.M. (1999) In situ swim- diversity. American Naturalist 100, 65–75. ming and settlement behaviour of larvae of an Parrish, J.K. (1999) Using behavior and ecology to exploit Indo-Pacific coral-reef fish, the Coral Trout (Pisces, schooling fishes. Environmental Biology of Fishes 55, Serranidae, Plectropomus leopardus). Marine Biology 157–81. 134, 51–64. Pauly, D., Christensen V., Froesse, R. and Palomares, Levin, P.S. (1993) Habitat structure, conspecific presence M.L. (2000) Fishing down aquatic food webs. American and spatial variation in the recruitment of a temperate Scientist 88, 46–51. reef fish. Oecologia 94, 176–85. Pennington, M. and Strømme, T. (1998) Surveys as a Ley, J.A., McIvor, C.C. and Montague, C.L. (1999) Fishes research tool for managing dynamic stocks. Fisheries in mangrove prop-root habitats of northeastern Florida Research 37, 97–106. Bay: distinct assemblages across an estuarine gradient. Pianka, E.R. (1966) Latitudinal gradients in species Estuarine Coastal and Shelf Science 48, 701–23. diversity: a review of concepts. American Naturalist Lindeman, K.C. and Snyder, D.B. (1999) Nearshore 100, 33–46. hardbottom fishes of southeast Florida and effects of Pihl, L. and Ulmestrand, M. (1993) Migration pattern habitat burial caused by dredging. Fisheries Bulletin of juvenile cod (Gadus morhua) on the Swedish west 97, 508–25. coast. ICES Journal of Marine Science 50, 63–70. Livingston, R.J. (1982) Trophic organization of fishes in Platell, M.E. and Potter, I.C. (1999) Partitioning of habitat a coastal seagrass system. Marine Ecology Progress and prey by abundant and similar-sized species of Series 7, 1–12. the Triglidae and Pempherididae (Teleostei) in coastal Longhurst, A.R. (ed.) (1981) Analysis of Marine Ecosys- waters. Estuarine Coastal and Shelf Science 48, tems. London: Academic Press. 235–52. Love, M.S. and Ebeling, A.W. (1978) Food and habitat of Polunin, N.V.C. (1996) Trophodynamics of reef fisheries three switch-feeding fishes in the kelp forests off Santa productivity. In: N.V.C. Polunin and C.M. Roberts Barbara, California. Fisheries Bulletin 76, 257–71. (eds) Reef Fisheries. New York: Chapman & Hall, pp. Mann, K.H. (1993) Physical oceanography, food chains, 113–35. and fish stocks: a review. ICES Journal of Marine Rhodes, K.L. (1998) Seasonal trends in epibenthic fish Science 50, 105–19. assemblages in the near-shore waters of the western Mann, K.H. and Lazier, J.R.N. (1996) Dynamics of Yellow Sea, Qingdao, People’s Republic of China. Marine Ecosystems: Biological–Physical Interactions Estuarine Coastal and Shelf Science 46, 629–43. in the Oceans, 2nd edn. Boston: Blackwell Scientific. Roberts, C.M. (1996) Settlement and beyond: population Marshall, N. (1985) Ecological sustainable yield (fish- regulation and community structure of reef fishes. In: eries potential) of coral reef areas, as related to N.V.C. Polunin and C.M. Roberts (eds) Reef Fisheries. physiographic features of coral reef environments. London: Chapman & Hall, pp. 86–112. Proceedings of the 5th International Coral Reef Robertson, D.R. and Allen, G.R. (1996) Zoogeography of Congress 5, 525–30. the shorefish fauna of Clipperton Atoll. Coral Reefs 15, McCormick, M.L. (1995) Fish feeding on mobile benthic 121–31. invertebrates: influence of spatial variability in habitat Rogers, S.I., Maxwell, D., Rijnsdorp, A.D., Damm, U. and associations. Marine Biology 121: 627–37. Vanhee, W. (1999) Fishing effects in northeast Atlantic Millar, R.B. and Fryer, R.J. (1999) Estimating the size- shelf seas: patterns in fishing effort, diversity and com- selection curves of towed gears, traps, nets and hooks. munity structure. IV. Can comparisons of species di- Reviews in Fish Biology and Fisheries 9, 89–116. versity be used to assess human impacts on demersal Moyle, P.B. and Cech, J.J. Jr (1996) Fishes: An Introduc- fish faunas? Fisheries Research 40, 135–52. tion to Ichthyology. Upper Saddle River, NJ: Prentice- Rose, G.A. (1993) Cod spawning on a migration highway Hall. in the north-west Atlantic. Nature 366, 458–60. Ecology of Marine Fish Communities 357

Rosenzweig, M.L. (1995) Species Diversity in Space and Sharp, G.D. (1987) Climate and fisheries: cause and effect Time. Cambridge, MA: Harvard University Press. or managing the long and short of it all. South African Ruzzante, D.E., Taggart, C.T. and Cook, D. (1999) A re- Journal of Marine Science 5, 811–38. view of the evidence for genetic structure of cod Sherman, K. (1994) Sustainability, biomass yields, and (Gadus morhua) populations in the NW Atlantic and health of coastal ecosystems: an ecological perspec- population affinities of larval cod off Newfoundland tive. Marine Ecology Progress Series 112, 277–301. and the Gulf of St. Lawrence. Fisheries Research 43, Shulman, M.J. (1985) Recruitment in coral reef fishes: ef- 79–97. fects of distribution of predators and shelter. Ecology Ruzzante, D.E., Wroblewski, J.S., Taggart, C.T., 66, 1056–66. Smedbol, R.K., Cook, D. and Goddard, S.V. (2000) Smith, S.V. (1978) Coral reef area and the contributions of Bay-scale population structure in coastal Atlantic cod reefs to processes and resources of the world’s oceans. in Labrador and Newfoundland, Canada. Journal of Nature 273, 225–6. Fish Biology 56, 431–47. Sogard, S.M., Powell, G.V.N. and Holmquist, J.G. (1989) Sadovy, Y., Rosario, A. and Roman, A. (1994) Reproduc- Fish utilization of shallow, seagrass-covered banks in tion in an aggregating grouper: the red hind, Epineph- Florida Bay I. Species composition and spatial hetero- elus guttatus. Environmental Biology of Fishes 41, geneity. Environmental Biology of Fishes 24, 53–65. 269–86. Spalding, M.D. and Grenfell, A.M. (1997) New estimates Sale, P.F. (1978) Reef fish and other vertebrates: a com- of global and regional coral reef areas. Coral Reefs 16, parison of social structure. In: E.S. Reese and F.J. 225–30. Lighter (eds) Contrasts in Behavior. New York: Wiley- Springer, V.G. (1982) Pacific Plate Biogeography with Interscience, pp. 313–46. Special Reference to Shorefishes. Smithsonian Contri- Sale, P.F. (1980) The ecology of fishes on coral reefs. butions to Zoology No. 36. Washington, DC: Smith- Oceanography and Marine Biology Annual Review sonian Institution Press. 18, 367–421. Stevens, G.C. (1989) The latitudinal gradient in geo- Sale, P.F. (1988) What coral reefs can teach us about ecol- graphical range: how so many species coexist in the ogy. Proceedings of the 6th International Coral Reef tropics. American Naturalist 133, 240–56. Symposium 1, 19–27. Stobutzki, I.C. (1998) Interspecific variation in sustained Sale, P.F. (1991a) Habitat structure and recruitment in swimming ability of late pelagic stage reef fish from coral reef fishes. In: S.S. Bell, E.D. McCoy and H.R. two families (Pomacentridae and Chaetodontidae). Mushinsky (eds) Habitat Structure: The Physical Coral Reefs 17, 111–19. Arrangement of Objects in Space. London: Chapman Stobutzki, I.C. and Bellwood, D.R. (1998) Nocturnal ori- & Hall, pp. 197–210. entation to reefs by late pelagic stage coral reef fishes. Sale, P.F. (1991b) The assembly of reef fish communities: Coral Reefs 17, 103–10. open non-equilibrial systems. In: P.F. Sale (ed.) The Swain, D.P., Chouinard, G.A. and Drinkwater, K.F. (1998) Ecology of Fishes on Coral Reefs. San Diego, CA: Acad- Seasonal variation in the habitat associations of emic Press, pp. 564–98. Atlantic cod (Gadus morhua) and American plaice Sale, P.F. (1996) Structure and dynamics of reef fish com- (Hippoglossoides platessoides) from the southern Gulf munities: a biogeographic comparison. In: M.L. Cody of St. Lawrence. Canadian Journal of Fisheries and and J.A. Smallwood (eds) Long-term Studies of Verte- Aquatic Sciences 55, 2548–61. brate Communities. San Diego, CA: Academic Press, Sweatman, H.P.A. (1988) Field evidence that settling pp. 73–97. coral reef fish larvae detect resident fishes using dis- Sale, P.F. (2002) The science we need to develop for more solved chemical cues. Journal of Experimental Marine effective management. In: P.F. Sale (ed.) Coral Reef Biology and Ecology 124, 163–74. Fishes: Dynamics and Diversity in a Complex Ecosys- Sweatman, H.P.A. and St John, J. (1990) Effects of selec- tem. San Diego, CA: Academic Press, pp. 361–76. tive settlement and of aggression by residents on distri- Sale, P.F. and Tolimieri, N. (2000) Density-dependence at bution of young recruits of two tropical damselfishes. some time and place? Oecologia 124, 166–71. Marine Biology 105, 247–52. Sale, P.F., Guy, J.A. and Steel, W.J. (1994) Ecological struc- Templeman, W. (1966) Marine resources of Newfound- ture of assemblages of coral reef fishes on isolated land. Bulletin of the Fisheries Research Board of patch reefs. Oecologia 98, 83–99. Canada 154, 1–170. Serchuk, F.M. and Wigley, S.E. (1992) Assessment and Thayer, G.W., Colby, D.R. and Hettler, W.F. (1987) Uti- management of the Georges Bank cod fishery: an his- lization of the reef mangrove prop root habitat by fishes torical review and evaluation. Journal of Northwest in south Florida. Marine Ecology Progress Series 35, Atlantic Fishery Science 13, 25–52. 25–38. 358 Chapter 16

Thresher, R.E. (1991) Geographic variability in the ecolo- in Hawaiian reef fishes. Environmental Biology of gy of coral reef fishes: evidence, evolution, and possible Fishes 18, 257–76. implications. In: P.F. Sale (ed.) The Ecology of Fishes Walters, C., Christensen V. and Pauly, D. (1997) Structur- on Coral Reefs. San Diego, CA: Academic Press, pp. ing dynamic models of exploited ecosystems from 401–36. trophic mass-balance assessments. Reviews in Fish Tolimieri, N. (1995) Effects of microhabitat characteris- Biology and Fisheries 7, 139–72. tics on the settlement and recruitment of a coral reef Warner, R.R. (1995) Large mating aggregations and daily fish at two spatial scales. Oecologia 102, 52–63. long-distance spawning migrations in the bluehead Tolimieri, N. (1998) Effects of substrata, resident con- wrasse, Thalassoma bifasciatum. Environmental specifics and damselfish on the settlement and recruit- Biology of Fishes 44, 337–45. ment of the stoplight parrotfish, Sparisoma viride. Watling, L. and Norse, E.A. (1998) Disturbance of the Environmental Biology of Fishes 53, 393–404. seabed by mobile fishing gear: a comparison to forest Tolimieri, N., Sale, P.F., Nemeth, R.S. and Gestring, K.B. clearcutting. Conservation Biology 12, 1180–97. (1998) Replenishment of populations of Caribbean reef Wellington, G.M. and Victor, B.C. (1985) El Nino fishes: are spatial patterns of recruitment consistent mass coral mortality: a test of resource limitation through time? Journal of Experimental Marine Biol- in a coral reef damselfish population. Oecologia 68, ogy and Ecology 230, 55–71. 15–19. Tupper, M. and Boutilier, R.G. (1997) Effects of habitat on Williams, D.McB. (1986) Temporal variation in the struc- settlement, growth, predation risk and survival of a ture of reef slope fish communities (central Great temperate reef fish. Marine Ecology Progress Series Barrier Reef): short-term effects of Acanthaster planci. 151, 225–36. Marine Ecology Progress Series 28, 157–64. Turner, S.J., Thrush, S.F., Hewitt, J.E., Cummings, V.J. Williams, D.McB. and Sale, P.F. (1981) Spatial and tem- and Furnell, G. (1999) Fishing impacts and the degrada- poral patterns of recruitment of juvenile coral reef tion or loss of habitat structure. Fisheries Management fishes to coral habitats within One Tree Lagoon, Great and Ecology 6, 401–20. Barrier Reef. Marine Biology 65, 245–53. Verity, P.G. and Smetacek, V. (1996) Organism life Wood, E.J.F. (1967) Microbiology of Oceans and Estuar- cycles, predation, and the structure of marine pelagic ies. Amsterdam: Elsevier. ecosystems. Marine Ecology Progress Series 130, 277– Woods, J.D. (1988) Scale, upwelling and primary produc- 93. tion. In: B.J. Rothschild (ed.) Towards a Theory on Victor, B.C. (1983) Recruitment and population dynam- Biological–Physical Interactions in the World Ocean. ics of a coral reef fish. Science 219, 419–20. London: Kluwer, pp. 7–38. Victor, B.C. (1986) Larval settlement and juvenile mor- Zeller, D.C. (1997) Home range and activity patterns of tality in a recruitment-limited coral reef fish popula- the coral trout Plectropomus leopardus (Serranidae). tion. Ecological Monographs 56, 145–60. Marine Ecology Progress Series 154, 65–77. Walsh, J.J. (1981) Shelf-sea ecosystems. In: A.R. Zeller, D.C. (1998) Spawning aggregations: patterns of Longhurst (ed.) Analysis of Marine Ecosystems. movement of coral trout Plectropomus leopardus London: Academic Press, pp. 159–96. (Serranidae) as determined by ultrasonic telemetry. Walsh, W.J. (1987) Patterns of recruitment and spawning Environmental Biology of Fishes 34, 357–77. 17 Interactions Between Fish, Parasites and Disease

I. BARBER AND R. POULIN

17.1 INTRODUCTION review all aspects of the subject. Our approach has been to concentrate on what we regard as currently Water provides the ideal medium for the survival, dynamic research fields, highlighting those that dispersal and proliferation of disease-causing or- examine the evolutionary and ecological conse- ganisms. Fish are particularly important hosts quences of parasite infections. We have attempted of parasites in aquatic ecosystems, harbouring to present our review in a manner that is of in- a wide variety of adult and immature forms and terest and use to both theoretical and empirical acting either as the sole host in a parasite’s life scientists and those involved in fisheries manage- cycle or as one in a series of hosts. Some parasites ment and, where possible, we have illustrated our may be responsible for acute, economically im- text with examples drawn from studies of ecologi- portant outbreaks of disease in exploited fish cal, fisheries or aquaculture relevance. However, populations or reduce productivity through nutri- many of the studies that have given greatest tional effects; others may be responsible for insight have utilized a small number of model chronic long-term changes in population struc- systems and our review inevitably reflects this ture. Infection-associated changes to host pheno- literature bias. Our approach means we have omit- type may influence the outcome of intraspecific or ted various topics perhaps expected in a more tradi- interspecific interactions and have important con- tional review of fish parasitology. Readers seeking sequences for individual performance, whether insight into the pathology and treatment of eco- measured by an ecologist in terms of evolutionary nomically important fish diseases are referred to fitness or by a fisheries manager as reductions in Roberts (1989) and Bruno and Poppe (1996). Simi- stock biomass. Finally, since some fish parasites larly, for specific aspects of the biology of the vari- are transmittable to humans, and others reduce ous parasite groups themselves, we refer the reader the market value of fish products by either spoil- to the modern and classic texts documented in ing host tissues or reducing the demand for fish Table 17.1. We have attempted to draw the reader’s as food, infections may have socioeconomic or attention to other specialist texts in other areas human health importance. There is indeed some where we feel our coverage has been necessarily aspect of fish parasitology that is relevant to every compromised. modern fish biologist. We begin by introducing the taxonomic di- versity of organisms parasitic in and on fish, their 17.1.1 Aims and scope of this chapter equally varied life cycles and the effect of different environmental parameters on the richness and Given the remarkable breadth of fish parasitology types of fish parasite communities. We continue research hinted at above, we are clearly unable to by discussing host–parasite coevolution and other 360 Chapter 17 evolutionary aspects of host–parasite relation- lans, leeches, certain copepod, branchiuran and ships, particularly examining how the effects that isopod crustaceans, the glochidia larvae of certain parasites have on host phenotype may influence molluscs and lampreys. References to classic and the ecology and ‘fitness’ of infected fish. We then modern texts dealing with the major groups of fish develop this individual-based approach to an ex- parasites are provided in Table 17.1. amination of the types of population-level effects parasite infections may have in natural habitats, 17.2.2 Parasite life-cycle variation and ask what may happen when such systems are upset by anthropogenic change. Turning our atten- Parasites show tremendous variation in the types tion to the socioeconomic and human health im- of life cycles they exhibit. A useful distinction can plications of fish parasites, we examine how the be made between those parasites with direct life real and perceived threat of fish-borne human in- cycles, which involve individuals of a single host fection may impact on fisheries and aquaculture, species but sometimes interspersed with a free- and then look at how parasite infections may be living stage, and those with indirect life cycles controlled in natural and artificial populations. We that utilize a series of different host species. Such a conclude our chapter by highlighting two recently functional distinction is not related perfectly to developed applications of fish parasitology; using parasite taxonomy, since closely related members infections as ecological markers and as pollution of a particular parasite group may exhibit either monitoring systems. type of life cycle. However, it is useful because it identifies the mode of transmission, an impor- tant determinant of the importance of specific 17.2 FISH PARASITE pathogens under particular situations. Knowledge DIVERSITY of the life cycle and the mode of transmission of any parasite is the key to effective control; how- 17.2.1 Important groups ever life-cycle details are not known for many para- of fish parasites sites. For example, Kudoa spp. are myxosporean parasites of marine fish that produce proteolytic In this review we consider as fish parasites all or- enzymes to break down muscle filaments, provid- ganisms that have a requirement to live on or in ing nutrients for the parasite and causing the fish hosts for at least part of their lives and which, characteristic ‘black tail’ condition. Because the during that time, have some negative influence parasite does not have direct access to the external on their fish hosts. We therefore include some environment, spores are only released follow- vertebrates (lampreys) but not egg predators or the ing the disintegration of muscle tissue. How the ‘sneaky’ conspecific males of some fish species parasites get into a new fish host is not known that are often described as ‘parasitizing’ the mate (Moser and Kent 1994); the life cycle could resources of other fish (Forsgren et al., Chapter 10, either be direct or involve an intermediate host, this volume). Parasite taxa may be conveniently, as do other myxosporean life cycles (Kent et al. though arbitrarily, separated into two major 2001). groups, based primarily on their size and mode of In this section we highlight life cycles from sev- multiplication. The microparasites of fish, which eral important parasite groups, drawing on specific are generally capable of asexual multiplication examples from species of ecological, economic or within or on a single fish host, comprise the evolutionary importance for which conclusive viruses, bacteria and fungi, as well as the flagellated life-cycle details have been determined. and ciliated protozoa; these are frequently im- portant pathogens of fish in culture (Sommerville Directly transmitted parasites 1998). The metazoan macroparasites of fishes include the monogenean, digenean and cestode Parasites with direct life cycles reproduce either by platyhelminth worms, nematodes, acanthocepha- sporulation, fission, budding, egg production or in Fish, Parasites and Disease 361

Table 17.1 Taxonomy and key references to the major groups of organisms parasitic on or in fishes.

Parasite group Representative taxa Reference

Microparasites Woo and Bruno (1999) Viruses Infectious pancreatic necrosis (IPN), viral hemorrhagic septicaemia (VHS), Wolf (1988) infectious haematopoietic necrosis (IHN), channel catfish virus (CCV) Bacteria Aeromonas, Pseudomonas, Yersinia, Flexibacter, Renibacterium Austin and Austin (1993) Fungi Ichthyophonus, Saprolegnia

Macroparasites Woo (1995) Protista Lom and Dyková (1992) Apicomplexa Cryptosporidium Microspora Glugea Myxozoa Kudoa, Myxobolus (Myxosoma) Kent et al. (2001) Ciliophora Ichthyophthirius, Trichodina Platyhelminthes Williams and Jones (1994) Monogenea Gyrodactylus, Dactylogyrus Dawes (1947) Cestoidea Diphyllobothrium, Bothriocephalus, Ligula, Eubothrium, Schistocephalus Khalil et al. (1994) Trematoda Diplostomum, Allocreadium, Cryptocotyle Dawes (1947) Nematoda Anisakis, Pseudoterranova, Contracaecum Moravec (1994) Acanthocephala Corynosoma, Neoechinorhynchus, Acanthocephalus, Pomphorhynchus Crompton and Nickol (1985) Mollusca Unionidae Glochidia larvae Annelida Hirudinea Piscicola, Hellobdella, Hemibdella Sawyer (1986) Arthropoda Copepoda Caligus, Lernaeocerca, Salmonicola, Thersitina Kabata (1979) Isopoda Lironeca Branchiura Argulus Metazoa Agnatha Petromyzon Hardisty and Potter (1971)

some cases viviparity (Sommerville 1998). Para- 1998). Here we present two illustrative examples site groups with members that exhibit direct life of direct life cycles. cycles include the viruses, bacteria, fungi, proto- zoans, microsporeans and ectoparasitic monoge- Ichthyophthirius multifiliis (‘white spot’ disease; neans, crustaceans and leeches. Even though such Ciliophora: Ichthyophthiriidae) Ichthyophthir- parasites only rarely cause problems in natural ius multifiliis, the causative agent of ‘white spot’, a populations, when outbreaks do occur they may common and pathogenic disease of fish in tem- have catastrophic effects on fish stocks (e.g. perate and tropical aquaculture, is a directly trans- Rahimian and Thulin 1996). Such parasites are mitted ciliate. The endoparasitic trophont stages, particularly important pathogens in aquaculture located within the epidermis of infected fish, grow as a consequence of the high stocking densities, and divide for a period of about 1 week before ac- reduced water quality and general stress levels tively leaving the host and encysting as a tomont of fish, which provide ideal opportunities for stage in the aquatic environment. Each tomont un- their transmission and proliferation (Sommerville dergoes binary fission for a period of up to 24h to 362 Chapter 17 produce up to 104 tomites, which differentiate into and lining of the digestive tract. Transformation motile, positively phototactic theront stages that into M. cerebralis spores within the fish takes 2–3 are infective to fish. Once a host is located, pene- months (El-Matbouli and Hoffmann 1989). tration of the epidermis is swift and is followed by growth to the mature trophont stage within 7 days Diplostomum spathaceum (Trematoda: Digenea) (Matthews 1994). Diplostomum spathaceum is a trematode that infects a wide range of freshwater fish when Caligus spp. (Crustacea: Copepoda: Siphonostom- free-swimming cercariae, released from para- atoida) These important parasites of marine fish- sitized lymnaeid snails, penetrate the fish’s skin. es begin life as non-parasitic, planktonic naupliar The parasites migrate to the eye of the fish host, stages that moult to form the first copepodid stage. where they occupy the lens, causing exophthalmia The copepodid is parasitic and must locate a fish (Chappell et al. 1994) and impairing vision to host, to which it clings using prehensile antennae the extent that host foraging success is reduced before moulting into a specialized chalimus stage. (Crowden and Broom 1980). The definitive hosts of This stage is followed by three more instar phases, the parasites are piscivorous birds, particularly which are all attached to the host by an invasive gulls (Laridae), which acquire the parasite after frontal filament. Two preadult stages follow, feeding on infected fish. High levels of infestation which detach from the frontal filament and be- may cause losses in aquaculture and lead to a come capable of movement over the fish’s body reduced catch rate by anglers in salmonid sport surface and of transfer between hosts, feeding on fisheries (Moody and Gaten 1982). epidermal tissue using specialized mouthparts (Schmidt and Roberts 1989). Schistocephalus solidus (Cestoda: Pseudophyl- lidea) This common parasite of lacustrine popu- lations of three-spined stickleback (Gasterosteus Indirectly transmitted parasites aculeatus) utilizes the fish as its second intermedi- Indirect life cycles are particularly common in ate host (Fig. 17.1). Fish acquire the infection after some myxosporeans, most digenean trematodes, ingesting procercoid parasites, harboured by cy- cestodes, acanthocephalans and nematodes, where clopoid copepods, which penetrate the fish’s intes- transmission to the fish host is generally via pre- tine and grow to a large size in the body cavity, dation events or skin penetration by free-living significantly distending the abdomen of the host as motile stages. Fish may be used as intermediate they do so (Fig. 17.2). The increased nutritional de- hosts by larval stages or as definitive hosts by adult mand of the parasite and its combined effects on parasites capable of sexual reproduction. Here we host appetite, locomotion and competitive ability outline basic life cycles of representative species of have severe consequences for the behaviour of these groups. infected sticklebacks, changes that have been intensively studied (reviewed by Milinski 1990; Myxobolus cerebralis (Myxozoa: Myxosporea) Barber and Huntingford 1995). Many behaviour Spores released into the aquatic environment changes are likely to facilitate transmission to when infected fish die and decompose, or are con- bird and mammal definitive hosts, which become sumed by predators or scavengers, are ingested by infected after feeding on infected fish. This life- tubificid oligochaete worms. After 3–4 months in cycle pattern is repeated for many other cestode the worm’s gut epithelium, these spores develop parasites of economic or zoonotic importance, into the infective ‘actinosporean’ phase. These are including Ligula intestinalis (Fig. 17.3) and released into the water and enter susceptible fish, Diphyllobothrium spp. mainly salmonids, through the epithelial cells of the skin, fins, the base of the gills in the buccal cav- Anisakis simplex (Nematoda: Anisakidae) This ity and elsewhere in the mouth, upper oesophagus and other members of the genus Anisakis are Fish, Parasites and Disease 363

Fig. 17.1 Birds harbouring the adult cestode in their intestine transmit the parasite eggs into the aquatic environment with their faeces (a) where they hatch, releasing free-swimming coracidia (b). If these coracidia are ingested by copepods (c) then they pass into the body cavity and develop to form the procercoid stage (pr). If infected copepods are ingested by a three-spined stickleback (d) then the parasites move to the body cavity of the fish, where they undergo rapid growth (e), developing into plerocercoids (pl). The life cycle is completed when fish-eating birds ingest sticklebacks harbouring plerocercoids (f), allowing development to the egg-producing adult form in the avian intestine. parasites that utilize marine fish as intermediate range, diet, local habitat use and size of the host hosts. Fish acquire the parasite following the in- fish. As such, the examination of parasite commu- gestion of infected crustaceans, whereupon the nities may also be used as indicators, or markers, to worm penetrates and invades the flesh of the give an insight into the biology and habits of the fish where it remains until eaten by the definitive host species, and this theme is taken up again in host, a marine mammal. It is likely that definitive later sections of this chapter. Here we consider the hosts may also become infected following the major factors affecting the abundance and com- consumption of infected crustaceans, meaning plexity of parasite communities in host fish. that the fish is a non-essential, or ‘paratenic’, host in the life cycle. The life cycle is completed Parasite communities of tropical vs. when parasite eggs pass out with the faeces of the temperate fishes definitive host and are ingested by susceptible crustaceans. Like those of free-living species, fish parasite communities in tropical aquatic habitats are typi- 17.2.3 Environmental parameters cally more diverse and rich than their temperate affecting parasite communities counterparts. Latitudinal variation in parasite community structure has been demonstrated in It is rare for individual fish to be infected with just congeneric freshwater eels from the UK and the one single parasite species in nature; it is far more Australian tropics, where the tropical hosts har- common to find fish harbouring a suite of infec- boured a far greater parasite species richness than tions. The factors governing the development of their temperate counterparts (Kennedy 1995). Two such parasite ‘communities’ within fish hosts are main hypotheses have traditionally been proposed complex but largely depend on the geographic to explain the increased diversity of communities 364 Chapter 17

(a)

Fig. 17.3 Two plerocercoids of the pseudophyllidean tapeworm Ligula intestinalis removed from the body cavity of a naturally infected European minnow.

parasites, increasing the opportunity for mutation (b) and subsequent speciation events, whereas the tropics also hold ancient species with equally ancient host–parasite associations. Poulin and Rohde (1997) attempted to separate these two fac- tors by comparing the species richness of ecto- parasitic metazoans on the heads of 111 marine fish species selected to span both temperate and tropical regions. After controlling for phylogeny, to ensure results are a real effect and not just phylo- genetic artefacts, they found that tropical species exhibited higher species richness and that richness was better explained by water temperature than (c) latitude alone, supporting the evolutionary rate explanation. Fig 17.2 Three-spined stickleback naturally infected with four Schistocephalus solidus plerocercoids (a) before, (b) during and (c) after dissection. Parasite communities of marine vs. freshwater fishes Similar variation in parasite community structure of both parasitic and free-living animals in tropical is typically found when comparing marine and habitats: the evolutionary time hypothesis and the freshwater fish species. Holmes (1990) compared evolutionary rate hypothesis, which relates to the findings of a study examining the parasite temperature (Rohde 1993). Increased parasite di- community structure of marine rockfishes with a versity in tropical fish may be explained by either. study on freshwater fish and found parasite com- Elevated non-seasonal temperatures lead to in- munities of the marine species to be more ‘com- creased growth and reproduction rates of hosts and plex’, defined as having more species for a given Fish, Parasites and Disease 365 number of individual parasites. Holmes attributed the difference in parasite community complexity 17.3 EVOLUTION OF to the greater mobility of, and habitat diversity HOST–PARASITE encountered by, marine fish hosts and the broader RELATIONSHIPS host specificity of marine gastrointestinal hel- minths. However, we are aware of no studies that 17.3.1 Host–parasite coevolution have examined the parasite fauna of groups with Cospeciation of fish and parasites? both marine and freshwater congeners, such as the sticklebacks or gobies, or of different populations From a historical perspective, the main question of single species found in the different habitats. regarding fish–parasite relationships has to do Both types of study would give useful comparative with their evolutionary origins. Is a parasite pre- data. sent on a fish because of an association with the an- cestors of the host that has been passed down through the host’s lineage or because of a relatively Other factors affecting parasite communities recent colonization event, that is, a switch from Within a particular geographic range, other another host lineage? The first scenario implies factors are likely to determine both the number that associated fish and parasite lineages cospeci- and species composition of parasite infections har- ate over evolutionary time such that their phyloge- boured by individual species. Habitat diversity and nies should be significantly congruent; host species richness both increase the likelihood switching events, on the other hand, would disrupt that the correct hosts and habitat requirements for this congruence. Monogenean and copepod ec- parasites with indirect life cycles or those with toparasites of fish are often assumed to have cospe- free-living stages will be met. The body size of a ciated with their hosts because of their high host particular fish host species may be important, specificity (Poulin 1992). The few phylogenetic since it affects host diet and determines both the analyses available tend to support this view (Gué- size of the space available for colonization and the gan and Agnèse 1991; Paterson and Poulin 1999). size of the ‘target’ the host presents to parasites Coevolutionary patterns between fish and endo- that infect via mobile free-living stages. Poulin and helminths have not yet been examined but host Rohde (1997) found that host body size correlated switching may have played an important role, as it closely with the number of individual ectopara- has in the coevolution of endohelminths and other sites but not the species diversity harboured by marine vertebrates (Hoberg 1995). marine fish. For parasites transmitted via the food Recent coevolutionary studies have shown that chain, diet may also be an important determinant parasite speciation and diversification can take of the types of parasites that infect fish hosts. A place on shorter time-scales. Molecular analyses study into the parasite fauna of two trophic have revealed that what were previously consid- morphs of Arctic charr Salvelinus alpinus in ered single species of fish parasites actually con- Loch Rannoch, Scotland, illustrates this. Parasite sisted of several distinct species or genetic entities communities of the pelagic morph, which feeds (Renaud and Gabrion 1988; De Meeüs et al. 1992). on planktonic prey, were dominated by copepod- Sympatric congeneric species, though highly simi- transmitted cestodes whereas those of the benthic lar and easily confused for one another, are often morph were dominated by acanthocephalans, produced rapidly through reproductive isolation transmitted via bottom-dwelling fauna, and following host switches. No doubt the use of diplostomatid trematodes, the cercaria of which molecular approaches will reveal other complexes are released from benthic snails (Dorucu et al. of cryptic species within ‘single’ parasite species 1995). known to occur in several fish species. 366 Chapter 17

heavily infected mates may have direct benefits, Parasites as agents of natural and such as the avoidance of potentially harmful dir- sexual selection in fish ectly transmitted contagious parasites or a reduced Three requirements are necessary for parasites to chance of passing microparasites to fry through act as selective agents: (i) their numbers must vary vertical transmission. Alternatively, in species among individual hosts in a population; (ii) varia- where male fish provide some degree of parental tion in parasite numbers among hosts must be as- care, females may avoid selecting mates that are sociated with variation in host fitness and (iii) the poor carers for eggs or young. In contrast Stott and trait under selection must either directly influence Poulin (1996) found that parasitized males may the number of parasites in an individual host or the invest heavily in nest defence. Evidence for indirect host’s ability to tolerate parasites, or the expres- benefits in fish, where choosy individuals are sion of the trait must depend on the number of proposed to benefit from mate choice by selecting parasites per host. Parasites in general meet these ‘good genes’ for their offspring, are scarce. However, conditions and can thus be direct agents of selec- recent studies do suggest that the offspring of bright tion (Goater and Holmes 1997). However, exam- male sticklebacks may have higher resistance to ples of parasite-mediated evolution in fish are few. S. solidus infections, attained at the expense of The striped bass (Morone saxatilis), introduced a slower growth (Barber et al. 2001; Hutchings, century ago from the east coast of North America Chapter 7, this volume). to the west coast, at first suffered severe pathology Parasites could ultimately be the reason why and mortality from infection by a west-coast fish male fish have evolved complex courtship dis- tapeworm. Today, bass from the west coast are plays, bright coloration or other forms of sexual more tolerant of the parasite than fish of east-coast ornamentation. At the same time, the proximate populations from which they derived (Sakanari debilitating effects of parasites can affect the sen- and Moser 1990), suggesting that they evolved to sory acuity of females or reduce the effort that cope with infections and illustrating how rapidly they invest in mate selection, making them less parasite-mediated natural selection can act in likely to discriminate among males differing in fish populations. Similarly, European minnows quality (Poulin 1994; López 1999). Thus the evi- (Phoxinus phoxinus) infected experimentally with dence available to date suggests that parasites can Diplostomum phoxini are less likely to die from mediate sexual selection either by driving it or the infection when challenged with sympatric slowing it down. parasites than if the parasites are cultured from Across species, Hamilton and Zuk (1982) pro- separate, genetically isolated host populations posed that host species exposed to strong selection (Ballabeni and Ward 1993). pressure from parasites should evolve sexual char- Much work has focused on the role of parasites acteristics that are more elaborate or pronounced as agents of sexual selection (Forsgren et al., than those of species incurring little pressures Chapter 10, this volume). Hamilton and Zuk (1982) from parasites. A comparative analysis of British hypothesized that male secondary sexual charac- freshwater fish revealed a positive correlation teristics serve as costly indicators of resistance to between sexual dichromatism and the number of parasites and that females should prefer males with parasite genera per fish species (Ward 1988). How- the most elaborate sexual characteristics. Several ever, a separate analysis of North American fresh- studies have confirmed that parasite load and the water fish showed that the relationship between expression of sexual traits in male fish are often neg- fish coloration and the number of parasite taxa atively correlated and that females prefer the exploiting a fish species is an artefact of sampling showiest males (reviewed by Kodric-Brown 1990); effort (Chandler and Cabana 1991). There is in sex-role-reversed pipefish, males avoid mating therefore no solid evidence at present that para- with heavily parasitized females (Rosenqvist and sites have played a role in sexual selection at the Johansson 1995). The avoidance and rejection of macroevolutionary level. Fish, Parasites and Disease 367

17.3.2 Evolution of host the consequences of parasite-induced changes in modification: why does it occur? host phenotype are probably limited to the most heavily infected individuals in a fish population. Parasites often induce various physiological and We discuss these consequences after reviewing the behavioural changes in their hosts. These changes mechanisms responsible for these alterations and have three possible causes (Poulin 1995). First, the specific types of parasite-induced phenotypic they may simply be non-adaptive coincidental alterations observed in fish. side-effects of infection. This is the most parsimo- nious explanation for simple changes in behav- 17.3.3 Infection-associated iour or performance, such as a reduction in behaviour change swimming speed in parasitized fish (Sprengel and Lüchtenberg 1991). Second, alterations in host Behaviour changes are frequently the most visible behaviour following infection may be the expres- manifestations of physiological or pathological ef- sion of parasite genes modifying the phenotype fects of parasite infections and, in an ecological and of the host. In other words, changes in hosts may evolutionary context, may be the most important be adaptive manipulations by the parasite that sublethal effects. Fish naturally perform a wide should facilitate the completion of its life variety of behaviours: on a daily basis they need to cycle. The complex phenotypic alterations in locate and compete for food and avoid predators sticklebacks induced by the cestode S. solidus (Juanes et al., Chapter 12, this volume; Krause appear to fall into this category; most observed et al., Chapter 13, this volume), whereas over changes seem likely to increase predation on longer time periods they need to find mates and infected sticklebacks by the predatory birds spawn successfully. The fulfilment of these re- that serve as definitive hosts for the parasite quirements may involve long-distance migrations (LoBue and Bell 1993; Tierney et al. 1993). (Mctcalfe et al., Chapter 8, this volume) or compe- Third, changes in infected hosts may represent tition over a territory or limited resources. If host responses to infection, aimed at eliminating infections interfere with host nutrition, physio- the parasites or compensating for their effects. logy or motor–sensory function, then changes to The greater risks taken by sticklebacks har- such behaviours may be expected. bouring S. solidus whilst foraging can be inter- There has been considerable recent interest preted as a fish strategy aimed at compensating in host behavioural change in fishes associated for the energy lost to the parasite (Milinski with parasite infections (see review by Barber et al. 1985; Godin and Sproul 1988). The stickleback– 2000), and we provide an overview of major studies cestode example illustrates that the above three that have revealed changes to specific behaviours explanations are not always mutually exclusive associated with parasite infections in Table 17.2. A and that they are difficult to disentangle from one common problem in many studies has been that another. of distinguishing between cause and effect (Poulin A single parasite larva may often be enough to 1998). Deviation from ‘normal’ behaviour patterns induce a measurable phenotypic modification in in naturally infected fish is insufficient to identify an invertebrate host. In fish, however, as in other parasitism as the causal mechanism, since an vertebrates, the influence of parasites on host individual already behaving differently as a result phenotype is usually proportional to the numbers of some other factor could be more susceptible to or biomass of parasites harboured by a fish. This infection. Yet despite providing the strongest is true whether the phenotypic changes are design for detecting infection-mediated behaviour reductions in swimming speed (Sprengel and change, studies in which infection status is con- Lüchtenberg 1991), visual acuity (Crowden and ferred on a random sample of an experimental Broom 1980) or antipredator behaviours (Poulin population are scarce, and in the majority of cases 1993; Lafferty and Morris 1996). This means that causality is yet to be ascertained. 368 Chapter 17

Mechanisms of behaviour change Even in cases where behavioural change in fish hosts has been convincingly attributed to infecting parasites, the mechanisms of behavioural modifi- cation have rarely been identified. Yet the elucida- tion of the mechanistic basis of behaviour change can provide important clues to the evolutionary background and likely ecological consequences of behaviour change. Here we outline the major types of mechanism that have been identified. Parasites that live in the intestine, grow in the body cavity or absorb nutrients from an invasive connection from the outside – as well as those that increase the metabolic cost of locomotion, stimu- late large immune responses or increase activity levels – may cause a significant drain on the energy reserves of host fish. Many behavioural changes undoubtedly arise from these energetic effects of infection, including altered foraging behaviour and appetite (Milinski 1990). Concurrent in- creases in respiratory demand, which may also re- sult from parasite-inflicted damage to gill surfaces, Fig. 17.4 Histological section showing two may lead fish to have a higher requirement for oxy- Diplostomum phoxini metacercariae in the brain tissue gen, leading to changes in the habitat preferences of a naturally infected European minnow. Metacercaris measure ~300mm ¥ 100mm. in order to exploit higher oxygen tensions. Certain infections have significant local patho- logical effects and for parasites that occupy sense 1977). These changes typically reduce body organs, muscles or neural tissue, local effects flexibility, locomotory speed and efficiency may be sufficient to bring about changes to host (Videler 1993). behaviour. Metacercariae of the trematode Dip- Parasites may also alter host behaviour through lostomum phoxini (Fig. 17.4) occupy specific brain more direct mechanisms, for example by produc- lobes of European minnows (Barber and Crompton ing host-modifying secretions. Alcohol and ketone 1997) and infection is associated with erratic waste products from encysted nematodes anaes- surface swimming of host fish (Rees 1955). thetize host musculature and reduce swimming Myxosporeans that destroy the cartilage of the performance (Ackman and Gjelstad 1975). The inner ear have similar effects on the swimming production of host hormone analogues is con- behaviour of salmonid hosts, causing ‘whirling sidered one of the strongest types of circumstan- disease’ (Markiw 1992), and reduced swimming tial evidence that behaviour change is an adaptive performance is also found in sheepshead minnows strategy by the parasite (Poulin 1998). Although infected with digeneans that encyst in the bulbus hormonal modification of host behaviour has been arteriosus and obstruct blood flow (Coleman demonstrated in invertebrate host–parasite sys- 1993). Parasites that grow to a large size and dis- tems, it has been difficult to detect in fish. Ligula tend the abdomens of host fish, such as cestode ple- intestinalis plerocercoids apparently castrate their rocercoids, increase host cross-sectional area and hosts through interactions with the pituitary– frictional drag (Rodewald and Foster 1998), as well gonadal axis of infected cyprinid fishes (Arme as causing atrophy of body musculature (Sweeting 1968), although the precise mechanism by which Fish, Parasites and Disease 369 this is achieved is yet to be discovered, despite ex- parasites (see Lafferty 1999 for review). Although tensive research effort (Williams et al. 1998). changes in antipredator behaviour of host fish associated with parasite infections have been demonstrated at every step of the predation cas- Types of behaviour change cade, from detection to identification, avoidance, Table 17.2 outlines the major studies that have susceptibility and capture, few studies have shown examined infection-associated changes in the that these behaviours increase susceptibility to behaviour of fish hosts, including changes in predators (see Section 17.3.4). Studies that demon- host foraging, swimming, antipredator, habitat strate rapid changes in host antipredator behaviour selection and reproductive behaviour. Here we around the period of parasite infectivity (Tierney summarize the major effects on each of these et al. 1993) are particularly interesting, since they behaviours. provide strong evidence for adaptive manipulation Fish harbouring parasite infections are frequ- by the parasite. For many small fishes inhabiting ently reported to be poor competitive foragers as a open water, shoaling is the primary antipredator result of impaired prey detection ability, poor prey defence strategy (Krause et al., Chapter 13, this vol- handling ability or impaired swimming perfor- ume), and infections have been shown to influence mance associated with infection. Such infections both shoal membership and performance within therefore present twin problems for host fish, since polarized schools (see Barber et al. 2000). The ex- as well as increasing nutritional demand, they tent to which antipredator behaviour is affected reduce the host’s ability to compete for food, and by infection may also depend on the host’s infected fish need to overcome this problem particular life-history strategy. For longer-lived (Milinski 1990). Demonstrated mechanisms by iteroparous species reduced antipredator perfor- which hosts mediate these disabilities include mance should be more costly for young fish, which modifying their time budget to increase the pro- have high expected future reproductive success, portion of time spent foraging, switching prey- than for older individuals. Young individuals size choice to exploit prey of lower quality and for should therefore invest more heavily in opposing which there is less competition, and foraging in the effects of infection. Poulin’s (1993) results are risky habitats avoided by non-parasitized fish. consistent with this theory, as they show that the Commonly reported effects of parasites on negative effects of trematodes on the antipredator swimming behaviour take the form of either responses of upland bullies (Gobiomorphus brevi- quantifiable reductions in swimming perfor- ceps) are more pronounced in 2+ and 3+ age-group mance, measured as maximum sustainable fish than in 1+ fish. velocity, burst speed or stamina, or changes in Common infection-associated changes in the swimming form. The latter changes are rarely spatiotemporal distribution, or habitat choice, quantified, comprising mainly descriptive or anec- of infected fish are the occupation of inshore dotal accounts of ‘erratic’ or otherwise conspicu- (cf. offshore) waters, their position in the water ous swimming movements. Both types of changes, column and preference for vegetation and cover which are unlikely to be mutually exclusive, may (Mittelbach, Chapter 11, this volume discusses be important in determining predation pressure non-parasite reasons for habitat switching). Al- on hosts, either by increasing their visibility to though such changes are often viewed as being predators or by impairing escape responses. host or parasite adaptations, care must be taken Since changes to antipredator behaviour poten- when interpreting the findings of field studies tially have immediate consequences for host showing differences in the habitats occupied by survival and for positively or negative effects on infected and uninfected fish. This is because fish parasite transmission, they have attracted a good occupying certain habitats or regions of a water deal of attention from evolutionary ecologists body may also be more or less likely to pick up interested in adaptive host manipulation by certain parasites. For example, minnows in Table 17.2 Selected studies documenting effects of parasites on host behaviour in teleost fishes.

Behaviour Specific Host species Parasite Parasite species Site of Marine Laboratory Direction Notes Reference category behaviour group infection (M)/Fresh (L)/Field (F) of change water (FW)

Foraging Time spent Three-spined Cestoda Schistocephalus Body cavity FW L Increase Giles (1987) foraging stickleback solidus Reactive Dace, three- Digenea Diplostomum Lens FW L Decrease Crowden and distance to spined spathaceum Broom (1980), prey stickleback Owen et al. (1993) Attack Dace Digenea Diplostomum Lens FW L Decrease Crowden and success spathaceum Broom (1980) Competitive Three-spined Cestoda Schistocephalus Body cavity FW L Decrease Under certain Barber and ability stickleback solidus conditions Ruxton (1998) Handling time Three-spined Cestoda Schistocephalus Body cavity FW L Increase Cunningham et al. stickleback solidus (1994) Prey choice Three-spined Cestoda Schistocephalus Body cavity FW L Change Preferred lower Milinski (1984) stickleback solidus quality (smaller) items Prey choice Three-spined Cestoda Schistocephalus Body cavity FW L Change Preferred higher Ranta (1995) stickleback solidus quality (larger) items Diet Three-spined Cestoda Schistocephalus Body cavity FW F Change Switched from Jakobsen et al. stickleback solidus planktonic (1988) cladocerans to benthos Diet Three-spined Cestoda Schistocephalus Body cavity FW F Change Less full Tierney (1994) stickleback solidus stomachs during winter and summer Swimming Ability to Sheepshead Digenea Ascocotyle Heart FW L Decrease Coleman (1993) sustain minnow pachycystis maximum velocity Swimming Smelt, eels Nematoda Pseudoterranova M L Decrease Sprengel and speed decipiens Lüchtenberg (1991) Burst swimming Coho, Digenea Nanophyetus FW L Decrease Butler and speed steelhead salmonicola Millemann (1971) Activity levels Guppy Digenea Diplostomum FW L Decrease Brassard et al. spathaceum (1982) Swimming Sockeye Myxosporidea Myxobolus FW L Decrease Moles and Heifetz speed arcticus (1998) Fatigue Sockeye Cestoda Eubothrium Intestine FW L Decrease Smith and distance salvelini Margolis (1970) Migration Sockeye Cestoda Eubothrium Intestine FW L Change Concentrated Smith (1973) performance salvelini towards the end of the run ‘Conspicuous European Digenea Diplostomum Brain FW L Change Infected fish Rees (1955) locomotion’ minnow phoxini exhibit ‘intermittent swimming movements’ ‘Conspicuous Various Myxosporidea Myxosoma Cranium FW L Change Infected fish Markiw (1992) locomotion’ salmonids cerebralis exhibit erratic circular swimming at water surface ‘Conspicuous Common Cestoda Ligula intestinalis Body cavity FW L Change Infected fish Dence (1958) locomotion’ shiner display ‘jerky’ or ‘sluggish’ swimming movements ‘Conspicuous Pejerrey Digenea Diplostomum Brain M L Change Fish harbouring Szidat (1969) locomotion’ mordax infective parasites impair movement Antipredator Risk avoidance Three-spined Cestoda Schistocephalus Body cavity FW L Decrease Infected fish Milinski (1990) behaviour stickleback solidus fed closer to predatory fish Three-spined Cestoda Schistocephalus Body cavity FW L Decrease Tierney et al. stickleback solidus (1993) Upland bully Trematoda Telogaster Body cavity FW L Decrease Reduction in Poulin (1993) opisthorchis antipredator behaviour increased with host age Table 17.2 Continued

Behaviour Specific Host species Parasite Parasite species Site of Marine Laboratory Direction Notes Reference category behaviour group infection (M)/Fresh (L)/Field (F) of change water (FW)

Risk avoidance Three-spined Cestoda Schistocephalus Body cavity FW L Decrease Infected fish Godin and Sproul stickleback solidus return quicker (1988) to food patch after being disturbed Social Shoal Three-spined Cestoda Schistocephalus Body cavity FW L Decrease Infected fish Barber et al. behaviour membership stickleback solidus left shoals (1995) quicker when food deprived Schooling European Cestoda Ligula intestinalis Body cavity FW L Change Increased Barber and behaviour minnow nearest Huntingford neighbour (1996) distance Schooling Killifish Trematoda Crassiphiala Skin FW L Change More often on Krause and Godin behaviour bulboglossa periphery (1994) Habitat Inshore water Pollack, Copepoda Lernaeacerca Gills M F Change Delayed Sproston and selection occupancy whiting branchialis seaward Hartley (1941) spring migration Inshore water Menhaden Copepoda Lernaeenicus Skin, gills M F Change Infected adults Guthrie and occupancy radiatus remained Kroger (1974) with juvenile Isopoda Olencira schools in praegustator estuaries Position in the Gudgeon, Cestoda Ligula intestinalis Body cavity FW F Change Infected fish Bean and Winfield water column roach swam higher (1992) in the water column Position in the Three-spined Cestoda Schistocephalus Body cavity FW L Change Infected fish LoBue and Bell water column stickleback solidus more (1993) buoyant and swam higher in water column Position in the Fathead Digenea Ornitodiplostoumum Brain FW L Change Infected fish Radabaugh water column minnow ptychocheilus swam closer to (1980) water surface Position in the Dace Digenea Diplostoumum Lens FW L Change Infected fish Crowden and water column spathaceum swam closer Broom (1980) to water surface Vegetation Three-spined Cestoda Schistocephalus Body cavity FW F Change Infected fish Jakobsen et al. preference stickleback solidus found more (1988) distant from vegetation Sexual Attractiveness Deep-snouted Digenea Cryptocotyle lingua Skin M L Decrease Males avoid Rosenqvist and behaviour to opposite pipefish mating with Johansson sex infected (1995) females Display rate Guppy Monogenea Gyrodactylus sp. Skin, gut FW L Decrease Infected males Kennedy et al. Nematoda Camallanus cottii displayed less (1987) frequently and were chosen as mates less often Egg care Upland bully Trematoda Telogaster Body cavity FW L & F Increase Infected males Stott and Poulin opisthorchis fanned more (1996) than uninfected males Mate inspection Upland bully Trematoda Telogaster Skin FW L Decrease Females spent Poulin (1994) opisthorchis less time inspecting males 374 Chapter 17 isolated estuarine sloughs have high digenean in parasitized fish (LoBue and Bell 1993; Poulin loads because of their proximity to snail interme- 1993; Krause and Godin 1994; Ness and Foster diate hosts (Coleman 1993) and the high frequency 1999), these findings suggest that phenotype- of eyeflukes in fish at specific sites in one lake has altering parasites can indirectly increase fish been explained by the local abundance of herons mortality by increasing their risk of predation. (Balling and Pfeiffer 1997). Unless it is possible The ecological and economic effects of pheno- to eliminate the effects of differential exposure type modification are not limited to enhanced to parasites caused by association with different predation. As reviewed above, infected fish may, habitats, then observational studies will generally for a variety of reasons, have reduced reproduc- be insufficient to demonstrate infection-mediated tive output. Parasite infection may also influ- habitat shifts. In addition, since habitat type may ence migration success. Since its introduction to determine the level of exposure to certain types Europe from Asia in the early 1980s, the nematode of parasite, fish may use habitat selection as a parasite Anguilicolla crassus has been causing primary strategy to avoid infection (see Section severe pathology and mortality in European eel 17.6). In this way, parasites may interfere with populations (Moravec 1992). Sublethal nematode the behaviour of (potential) hosts without actually infections reduce haematocrit and stamina of infecting them. eel hosts and cause important reductions in swim- Changes to sexual and reproductive behaviour ming ability (Sprengel and Lüchtenberg 1991), of parasitized fish has been well documented (see possibly making them less successful at complet- Kodric-Brown 1990; Table 17.2; Forsgren et al., ing their extensive migration. Considering that Chapter 10, this volume) and may stem from either helminth infections can also impair orientation reduced gonadal development (Pampoulie et al. in migrating fish (Garnick and Margolis 1990), it is 1999) or more subtle effects on sexual ‘attractive- clear that parasite-induced changes in host behav- ness’, parental ability or the discriminatory ability iour or physiology can have profound impacts on of the choosing sex. fish populations.

17.3.4 Ecological consequences 17.4 EFFECTS OF of host manipulation PARASITES ON FISH Whatever the phenotypic trait modified by POPULATION ECOLOGY parasites, and whether or not the modification is adaptive for host or parasite, the consequences for 17.4.1 Patterns of parasite fish populations may be important. In the majority distribution amongst of cases, the potential population effects of host populations even well-documented changes in fish phenotype caused by parasites remain unknown. What evi- The level of parasite infection can be quantified in dence there is, however, suggests that the influ- terms of the numbers of parasites of a particular ence of parasites can be substantial. Lafferty and species harboured by an individual (‘intensity’) or Morris (1996) demonstrated in the field that fish by the proportion of a host population that are in- parasitized by larval trematodes behave differently fected (‘prevalence’). To understand the potential and are much more susceptible to bird predation influence of a particular parasite on host popula- than unparasitized conspecifics. Other studies tion ecology, it is important that the specific have also found that parasitized fish are more pattern of infection within the host population vulnerable to predators than healthy ones (van is recognized. In natural host populations, it is Dobben 1952; Brassard et al. 1982). Combined generally the case that parasites are not distributed with the results of studies that report decreased equally amongst the population of hosts. Instead, antipredator responses or greater conspicuousness macroparasites typically exhibit an aggregated dis- Fish, Parasites and Disease 375 tribution, with most host individuals harbouring second outbreak typifies the destructive and op- low numbers of parasites and a small proportion of portunistic nature of such parasites in small the host population harbouring a disproportionate lowland lakes, where intense year-round avian load. The mechanisms leading to parasite overdis- predation on the host population ensures the persion may involve individual-level variation in parasite’s rapid transmission following a chance immunity or exposure to infective parasite stages, introduction to the system. Interestingly, the pat- or may result from stochastic processes (Poulin terns in Slapton Ley are very different to those in a 1998). lacustrine population of minnows infected with For longer-lived parasites, infection levels fre- Ligula in the highlands of Scotland, where the quently covary with host age, since parasites are parasite persisted at a low level for at least 10 years acquired over the lifespan of the fish. Body size, (I. Barber, unpublished data). The differences are which may covary with age, may additionally be probably attributable to the vastly differing preda- important if larger hosts provide a bigger ‘target tion regimes at the two sites: transmission of the area’ for infecting parasites. The combination of parasite at the highland site by gulls, the only pis- the two mechanisms may account for the fre- civorous birds of note at this site, is restricted to quent disproportionately high burdens harboured their presence during a short breeding period in by large host individuals. In temperate regions, or early summer. Lake size may also be important where hosts undertake regular migrations be- in determining the stability of infections, and in tween habitats differing in their intrinsic risk of larger lakes L. intestinalis may reach a state of parasite infection, there may additionally be a sea- equilibrium (Kennedy 1985). sonal effect on parasite load. However, such documented effects of parasites on natural fish populations are relatively few, and 17.4.2 Effects of parasites on host generally our best indications of population-level population structure effects of parasites are various types of circum- stantial evidence. For example, the small number The medium and long-term consequences of of extremely heavily infected fish predicted from parasite infections for fish populations are of great mathematical models of parasite distributions are interest to fisheries managers and parasito- often completely missing from host populations, logists alike. One of the most extensive long-term suggesting that they have been removed by preda- studies of host–parasite population change in a tors or have died naturally. This truncates the ex- natural system has been that of the population pected negative binomial distribution. Another consequences of Ligula intestinalis infection of clue is given by the absence of large heavily in- the roach population in Slapton Ley, a freshwater fected individuals from natural populations in lake in Devon, south west England (summarized field surveys; again, the implication is that these in Kennedy et al. 1994). Ligula intestinalis was fish, which may carry the majority of the parasite recorded for the first time in the roach population population, are not present because they have been in 1973, its introduction coinciding with the ar- removed from the population by predators (Poulin rival and successful breeding of great-crested 1998). grebes Podiceps cristatus, major definitive hosts of the parasite. The parasite spread rapidly through 17.4.3 Parasites and conservation: the roach population, reaching a prevalence of what happens when natural 28% 2 years later and causing a crash in the host systems are upset? population. Ligula intestinalis became gradually less prevalent and eventually disappeared from Natural and anthropogenic change may affect fish the population by 1985. The parasite was detected parasites, and hence diseases, in a number of ways. again at the site in 1990, and 12 months later the Firstly, parasites that have free-living stages will prevalence of the parasite had risen to 71%. This be susceptible to the same ecological parameters 376 Chapter 17 that control the distribution and abundance of wide host specificity for intermediate hosts. The free-living animals. Secondly, their reliance on economically important eel parasite Anguillicola hosts means that they will be affected by changes crassus (see above) is ideally equipped for to host populations. However, there is little direct colonization (Kennedy 1994). Although it has an evidence linking environmental changes to fish indirect life cycle, a wide range of copepod species parasites, and most evidence is circumstantial. are utilized as intermediate hosts and the para- site can also use a multitude of fish species as paratenic hosts. The parasite species need not al- Environmental impact on established parasites ways be completely novel to have such serious The consequences of environmental change, ei- consequences; a novel strain of an existing parasite ther long term or acute, natural or anthropogenic, may also be highly pathogenic in the ‘wrong’ host is an important area for research in parasitology population. Fish populations that have been sepa- (Hominick and Chappell 1993). Transmission be- rated for significant periods of time may support tween hosts is the most important rate-limiting highly coevolved populations of the same parasite step in all parasite life cycles, and both natural and species that are significantly more pathogenic than anthropogenic environmental changes that affect local coevolved strains (Ballabeni and Ward 1993). the rate at which parasites are transmitted poten- One of the most damaging outbreaks of a fish tially have massive impact on fish populations. For parasite in Europe in recent decades has resulted example, many parasites in temperate ecosystems from the introduction of Gyrodactylus salaris into that have free-living stages in their life cycles have the salmonid-rich rivers of northern Norway. This a lower temperature below which transmission monogenean parasite was first documented in 1973 does not occur (Chubb 1976). Where natural water and the subsequent decimation of salmonid stocks temperatures are artificially raised as the result of in these rivers over the following years led to the thermal pollution resulting from the input of cool- parasite being deemed a notifiable disease in 1983. ing water from power plants, such thermal limits Controversial treatment of infected rivers with may be constantly exceeded and infection possible rotenone in 1981 was successful in exterminating year round, leading to increased parasite loads the parasite locally (Johnsen and Jensen 1991). The in host fish (Högland and Thulin 1990). The parasite is thought to have been introduced with importance of long-term global climate change in stocked fish from infected hatcheries, possibly this respect is particularly relevant. Other types of from eastern Scandinavia where local salmonid pollution, such as nutrient loading resulting populations are resistant to the parasite (Bakke et al. from the input of sewage effluent, may alter com- 1990). munities of invertebrates, which often act as inter- mediate hosts, and of predators, which are often definitive hosts of fish parasites. Changes to such 17.5 SOCIOECONOMIC communities have the potential for knock-on ef- AND HUMAN HEALTH fects on fish parasites (see Section 17.7.2). IMPLICATIONS OF FISH PARASITES Anthropogenic parasite introductions 17.5.1 Economic importance of Introduction of a novel parasite species into a parasites in fisheries and aquaculture naïve population may have especially serious con- sequences for the fish hosts. Kennedy (1994) dis- There are four main ways in which parasites may cusses the biology of parasite introductions and impact on economically important fisheries identifies several characteristics of successful ‘pio- (Sinderman 1990). Firstly, direct mortality associ- neering’, or colonizing, parasites, including the ex- ated with highly pathogenic infections, or indirect istence of a life cycle that is direct or, if indirect, has mortality resulting from increased levels of preda- Fish, Parasites and Disease 377 tion on exploited stocks, may reduce the numbers may serve as a substrate for ectoparasitic proto- of fish available in the harvestable stock. Parasitic zoans. Certain discrepancies in the importance of fungi (Ichthyophonus) have been associated various groups between different types of systems with recent mass mortalities of economically are evident; for example, the microsporideans are important species (Rahimian and Thulin 1996). more important pathogens in mariculture than in Secondly, nutritionally demanding parasites or freshwater systems. However, indirectly transmit- those that impair foraging behaviour may cause ted parasites, such as digenean trematodes, may poor growth performance. Chronic low-level ces- also become problematic in aquaculture because tode infections of salmonids in aquaculture lead invertebrate populations, which are potential in- to economically significant reduced body weight termediate hosts, build up quickly in fish farms in at harvest (Bristow and Berland 1991), as do mi- response to nutrient availability. The availability of crosporidean infections (Hauck 1984) (but see definitive hosts also provides ample possibility for Arnott et al. 2000 for a demonstration of parasite- such infections to become established (Kennedy induced gigantism in fishes). Kudoa thyrsites has 1994), and predators attracted to the farm poten- an intensity-dependent effect on muscle texture in tially facilitate the completion of complex parasite Atlantic salmon Salmo salar, with heavily in- life cycles. The clinostomatid and heterophyid fected fish exhibiting ‘soft flesh’ disease (St-Hilaire trematodes in particular cause problems in warm- et al. 1997), which reduces market value and host water aquaculture because of their large size and survival. Thirdly, consumer aversion to visibly infectivity to humans respectively. detectable infections can increase wastage and processing costs and may severely reduce a fish- 17.5.2 Human health ery’s viability. Consumer rejection may be based implications of fish parasites on the presence of the parasites themselves, which in some cases may present a real human health risk The risks of infection with fish-borne parasites (see Table 17.3); more frequently it is based on present a potential or realized threat to the damage caused to host musculature, which re- health of human consumers particularly in soci- duces palatability, aesthetic quality and hence eties where the consumption of raw, undercooked marketable value (for marketing issues see Young or cold-smoked or lightly salted fish is traditional. and Muir, Chapter 3, Volume 2). This is particu- In their review of human helminths, Coombs larly problematic for aquaculture species, which and Crompton (1991) describe 76 registered are very often long-lived, high-value products. Fi- helminth pathogens of Homo sapiens that in- nally, changes in the size or structure of an infected volve fish in their life cycles, the majority of host population may have consequences for the which are acquired by humans following the con- performance of other species in the community sumption of infected fish (Table 17.3). Sinderman (some of which may also have commercial value). (1990) highlights three major groups of marine These mechanisms are likely to be complex but fish parasites with demonstrated public health may be positive, for example if they result in significance: the anisakid nematodes, the diphyl- emancipation from predation or competition, lobothriid cestodes and the heterophyid trema- or negative, for example if the parasite exhibits todes. These three groups of parasites give rise to density-dependent host switching. major diseases. Over 1000 cases of anisakiasis, The microparasites are generally the most im- resulting from the invasion of the wall of the portant parasites in aquaculture (Sommerville human digestive tract by Anisakis or Pseudoterra- 1998), since high stocking densities mean that para- nova sp. nematodes, were reported between 1969 sites with direct life cycles have a greater likelihood and 1977 in Japan, whereas diphyllobothriasis, of finding a host. Inevitable reductions in water resulting from the ingestion of infective Diphyl- quality may also irritate sensitive external fish sur- lobothrium sp. cestode plerocercoids, is an impor- faces, and the subsequent oversecretion of mucus tant disease in Asia. 378 Chapter 17 penetration Europe distribution route a Paratenic Geographic Infection ) a , ) Asia Ingestion ) Asia Ingestion Mugil , , ) Misgurnus Barbus Pelteobargus , Clarias Perca , ) ) Cyprinus including Tilapia Perca Acheilognathus Mogarnda a host birds host intermediate host Humans Snailincludingspecies (FW Fish Asia Ingestion Humans SnailspeciesBW and (FW Fish Africa Ingestion Piscivorous Snailspecies)(FW Fish Humans, Europe Skin Humans Snailincludingspecies (FW Fish Asia Ingestion Humans Snail(including Fish Europe Asia, Ingestion Humans Snail(including Fish Humans Snail(including Fish HumansHumans ?Humans SnailHumans SnailHumans(Salmonidae) Fish ?species)(FW Humans ?Fish Humansspecies)(FW Fish SnailHumans SnailHumans ?SnailHumans Snail(Cyprinidae) Fish (Cyprinidae) Fish Snail(Cyprinidae) Fish ? ?FishAmerica N. Asia(Cyprinidae) Fish species)(FW Fish Europe Ingestion (Cyprinidae) Fish Ingestion America, N. Ingestion America N. Europe Ingestion Ingestion Asia Asia Ingestion America C. Europe Ingestion Ingestion Ingestion Ingestion salmonicola schikhobalowi truncatum Clinostomum complanatum Clinostomum Prohemistomum vivax Prohemistomum Diplostomum spathaceum Diplostomum Echinochasmus japonicus Echinochasmus Echinochasmus perfoliatus Echinochasmus Echinostoma hortense Echinostoma Echinostoma malayanum Echinostoma Psilorchis hominis Psilorchis salmonicola Nanophyetus salmonicola Nanophyetus albidus Metorchis conjunctus Metorchis felineus Opisthorchis guayaquilensis Opisthorchis sinensis Opisthorchis viverrini Opisthorchis Pseudoamphistomum Parasites of humans that are acquired from fish or that otherwise involve fish in their life cycle (Source: data from Coombs and Table 17.3 Table Crompton 1991). taxonParasite SpeciesTrematoda Digenea Clinostomatidae Definitive Firsthost intermediate Second Cathycotylidae Diplostomatidae Echinostomatidae Psilostomatidae Nanophyetidae Opisthorchidae Fish, Parasites and Disease 379 America, Asia America, ) Asia Ingestion , , )))Asia Africa, Asia Africa, Ingestion N. Africa, Ingestion Ingestion , Asia Ingestion ) ) ,Asia Africa, Ingestion ) Europe , Asia Ingestion Liza Liza , , , , ) Mugil Mugil Puntius Puntius Puntius Barbus Mugil Butis Glossamia ) , , Mugil Aphanius , , Solea ) , ) ) Mugil Tilapia Acanthogobius Lateolabrax Acanthogobius Lateolabrax Mugil Chanos Acanthogobius Mugil Acanthogobius Dallia HumansHumansHumans SnailHumans SnailHumans SnailHumans SnailHumansspecies)(FW Fish SnailHumans(Cyprinidae) Fish Snailspecies)BW and (FW Fish ?SnailHumansspecies)BW and (M Fish Snail(including Fish (including Fish Snail ?Fish(including Fish AsiaAmerica N. (including Fish Asia Ingestion Ingestion Ingestion Ingestion Asia ?Ingestion Humans Snail(including Fish Humans ?Snail(including Fish Asia Ingestion Humans Snail(including Fish Asia Ingestion HumansHumansHumans Snail Snail Snail Fish(Cyprinidae) Fish (including Fish Asia Asia Ingestion Ingestion HumansHumans Snail Snail(including Fish (including Fish Asia Ingestion Humans Snail(including Fish America N. Ingestion Humans SnailHumansHumansHumans ?Copepodspecies)(BW Fish ?Copepod ?Copepodspecies)(M ?Fish species)(M ?Fish includingspecies (FW Fish Asia Asia Ingestion Ingestion Ingestion Ingestion Apophallus donicus Apophallus armatus Centrocestus formosanus Centrocestus lingua Cryptocotyle pumilio Haplorchis taichui Haplorchis vanissimus Haplorchis yokogawai Haplorchis dispar Heterophyes Heterophyes heterophyes Heterophyes Heterophyes nocens Heterophyes Heterophyopsis continua Heterophyopsis Metagonimus minutis Metagonimus yokogawai Metagonimus calderoni Procerovum Procerovum varium Procerovum summa Pygidiopsis Stellantchasmus falcatus Stellantchasmus Stictodora fuscata Stictodora Diphyllobothrium cameroni Diphyllobothrium chordatum Diphyllobothrium dalliae Diphyllobothrium Heterophyidae Cestoidea yllidea Pseudoph Diphyllobothriidae 380 Chapter 17 infected copepod) distribution route a Paratenic Geographic Infection a ) ) Ingestion ) Ingestion ) ) Ingestion ) ) Ingestion Coregonus Sardinops Megalocottus , , Scomberomorus , Eugraulis Elinginus Gasterosteus Onchrynchus Sarda Oncohrynchus host* animals (including dog) and cat Humans ?Copepod(including Fish HumansHumans ?CopepodHumans ?HumansHumans ?Fish ?Humans ? CopepodHumansincluding ?species (M Fish ?Fish Copepod ?Fish ?Fishincludingspecies (FW Fish includingspecies (M Fish Asia Asia Europe Ingestion America N. Ingestion Asia Europe Asia host intermediate host HumansHumansHumans Copepod ?CopepodHumans ?CopepodHumansspecies)(FW Fish Humansspecies)(M ?Fish ?Copepod ?Fish ?Copepod ?CopepodHumans ?FishCarnivorousspecies)(M ?Fish including(Salmonidae, Fish Copepod CopepodHumansspecies)(FW species)Fish (FW Humans; Fish Humans AsiaHumans ?CopepodHumans ?Copepod ?CopepodHumans ?Fish Asia Ingestion ?CopepodHumans ?Fish Asia Ingestion Humans(Salmonidae) Fish ?CopepodHumansincluding species (M Fish Ingestion ?Copepod ?Copepod ?Fish Ingestion Copepod(of Ingestion Ingestion ?Fish ?Fishincluding (Salmonidae, Fish Ingestion Ingestion Asia Ingestion Asia Ingestion Ingestion Ingestion Ingestion Ingestion erinaceieuropaei Diphyllobothrium lancelatum Diphyllobothrium Diphyllobothrium yonagoensis Diphyllobothrium balaenopterae Diplogomoporus branni Diplogomoporus fukuokaensis Diplogomoporus intestinalis Ligula anthocephalus Pyramicocephalus solidus Schistocephalus Diphyllobothrium dendriticum Diphyllobothrium elegans Diphyllobothrium Diphyllobothrium giljacicum Diphyllobothrium hians Diphyllobothrium klebanorskii Diphyllobothrium latum Diphyllobothrium mansoni Diphyllobothrium minus Diphyllobothrium nenzi Diphyllobothrium nihonkaiense Diphyllobothrium pacificum Diphyllobothrium scoticum Diphyllobothrium skrjabeni Diphyllobothrium tingussicum Diphyllobothrium ursi Diphyllobothrium Continued Table 17.2 Table taxonParasite Species Definitive Firsthost intermediate Second Fish, Parasites and Disease 381 Europe Asia America, ) ) species) species including Fundulus species including Ameirus ) Misgurnus ) crustaceanscrustaceans species) crustacea species) crustaceaamphipods species) species) M and BW species, including Hypseleotris crustaceans species) America, (Humansaccidental) copepods HumansHumans Euphausiid MarineHumansHumanscopepod FW Humanscopepod FW (Salmonidae) Fish includingspecies (FW Fish copepod FW Humansspecies)(FW (M Fish Fish Marine EuropeHumans(M Fish AsiaAsiaAmerica, ?Marine N. Asia Ingestion Ingestion C. ?Ingestion America, N. ?Ingestion Ingestion (M ?Fish America N. (BW Fish Ingestion America N. ?Ingestion HumansHumans Humans/fish?and(FW Fish Humans EuphausiidAmerica N. Asia Africa, Ingestion (M Fish Ingestion N. Asia, Ingestion Humans Oligochaete(FW Fish Ingestion sp. Humans Oligochaete(FW Fish America N. Ingestion sp. Humans ?Marine(M ?Fish Asia ?Ingestion sp. Fish MarineAmerica N. wound Via Contracaecum osculatum Contracaecum decipiens Pseudoterranova Philometra doleresi Gnathostoma hispidum Gnathostoma spinigerum Gnathostoma rauschii Acanthocephalus Bolbosoma strumosum Corynosoma Trichinella nativa Trichinella (Capillaria) Aonchotheca philappinensis (marina) simplex Anisakis Dioctophyma renale Dioctophyma Eustrongyloides FW, freshwater; BW, brackish water; M, marine. *? denotes non-confirmed or unknown host. freshwater; BW, FW, Dracunculidae Gnathostomatidae phala Acanthoce Echinorhynchidae Polymorphidae Trichinellidae Trichuridae Anisakidae Nematoda Dioctophymatidae a 382 Chapter 17

selves of sources of irritation such as skin par- 17.6 CONTROLLING asites (Wyman and Walters-Wyman 1985). Clean- PARASITE INFECTIONS ing symbioses represent a more sophisticated ectoparasite control method (see Losey 1987; 17.6.1 Natural parasite Poulin and Grutter 1996; Côté 2000 for reviews). control mechanisms This interaction involves the removal of ecto- parasites, mucus or diseased tissues from the Fish can adapt to tolerate infections. Tolerance is surfaces of fish by a cleaning organism. Cer- defined as minimizing the fitness effects of the tain crustaceans and fish, especially the wrasse parasite infection. However, two other strategies Labroides spp., are specialized cleaners of a wide are available for limiting parasite load: avoiding range of fish species; cleaning is also sometimes parasites or eliminating them following infection. observed between conspecific fish (Sikkel 1986). Several behavioural mechanisms can serve to de- The benefits of cleaning for fish are not always crease fish exposure to the infective stages of clear. Despite the large numbers of ectoparasites parasites. For instance, sticklebacks can avoid ingested by cleaners, experiments in which clean- certain microhabitats or form large shoals to de- ers were removed from reefs have revealed no crease their risk of being infected by ectoparasitic changes in infection levels by ectoparasites or no branchiuran crustaceans (Poulin and FitzGerald sign of fish emigrating to other reef patches (see 1989a,b). Parasites, or more correctly the threat of Poulin and Grutter 1996). The latest of these field infection they pose, also have the potential to alter experiments is the only one suggesting that the prey selection if fish have evolved to avoid either action of cleaners benefits fish: the action of prey items that contain infective parasite stages or cleaner wrasses led to a decrease in the numbers whole prey taxa, if their consumption is poten- of ectoparasitic gnathiid isopods per caged fish tially associated with subsequent infection. Few relative to fish caged on reefs without cleaners studies examining behavioural avoidance in fish (Grutter 1999). have been undertaken, but the results of Wedekind and Milinski (1996) suggest that although they 17.6.2 Biological control of are identifiable through altered behaviour, cope- parasites in fish farms pods infected with S. solidus are not avoided by foraging sticklebacks. Cleaner fish have recently been used as an alterna- Once infected, fish rely mainly on their im- tive to chemicals in order to control ectoparasite mune system to combat pathogens and parasites. infections on cultured fish. Specialized and/or Although its intricacies are not as well known as potential cleaner fish can reduce the number of those of the mammalian system, the fish immune monogeneans or copepods per fish when intro- system appears capable of controlling infections duced in tanks or sea cages containing cultured by both microorganisms and metazoan parasites fish such as tilapia or Atlantic salmon (Cowell (Iwama and Nakanishi 1996). Antibodies against et al. 1993; Costello 1996). Temperate wrasses not endoparasitic worms have been found in diverse known as specialized cleaners, but related to the fish taxa (Harris 1972; Sharp et al. 1989), and both tropical genus Labroides, show much potential for natural and acquired resistance against monoge- the biological control of sea lice (caligid copepods) neans has been demonstrated in guppies and on Atlantic salmon. A new inshore fishery has salmonids (Madhavi and Anderson 1985; Bakke recently appeared to supply wrasses to salmon et al. 1990). Much of the regulation of parasite farms. However this has other potential environ- numbers in wild fish may be due to immune re- mental implications, and wrasses themselves sponses (Lysne et al. 1997). are now being cultured to meet the demand Other modes of regulation can be used against (see Costello 1996). If these supply problems are ectoparasites. For example, fish are known to rub solved, there is no doubt that the tendencies of their body against hard substrates to rid them- certain fish to clean, and of most other fish to pose Fish, Parasites and Disease 383 for cleaners, can be further exploited for parasite ing results obtained using genetic markers, control. meristic data or any other type of biological indicator. 17.6.3 Chemical control, vaccine In recent years the use of parasites as ecological development and other approaches markers has been particularly important for the stock separation of commercially important fish Reactive treatment or prophylaxis of disease was species and the elucidation of recruitment migra- the primary mechanism of dealing with parasite tions of fish from nursery grounds to feeding and outbreaks, and the preventative use of chemical spawning grounds (Moser 1991; Williams et al. agents has been the typical approach taken to dis- 1992; Metcalfe et al., Chapter 8, this volume). The ease in aquaculture. The development of vaccines spatial scale over which infection levels by one against fish parasitic diseases was not considered parasite species in a fish species show significant feasible until very recently, yet it is now an impor- variation can be relatively small, and allows other tant industry. Vaccine development is a complex uses for parasites as ecological markers. For in- new field, the specific details of which lie beyond stance, on coral reefs, fish from the reef flat can the scope of this chapter; we refer the interested have significantly different ectoparasite and endo- reader to recent reviews by Evelyn (1997) and parasite loads compared with conspecifics from Vinitnantharat et al. (1999). the reef slope just a few hundred metres away Low-tech solutions may also be effective in (Rigby et al. 1997). Parasites can therefore shed controlling outbreaks of indirectly transmitted light on the limited movements of fish and on their parasites, principally by removing or controlling local population structure. However, the fact that intermediate hosts. Field and Irwin (1994) describe some parasites have the potential to alter host how a management strategy of increased water behaviour, including swimming performance flow and weed-bed, algae and snail removal at a and migratory patterns (see Table 17.2), should be fish farm in Northern Ireland reduced the preva- taken into account when identifying ideal indica- lence of Diplostomum spathaceum in rainbow tor parasites. trout from 100% to 20%. Other uses of parasites as biological tags in- clude the determination of fish diets and spa- tiotemporal feeding patterns, as well as the 17.7 RECENT reconstruction of food webs (Williams et al. 1992; APPLICATIONS OF Marcogliese and Cone 1997; Polunin and Pinnegar, FISH PARASITOLOGY Chapter 14, this volume describe other methods). Such studies focus on parasites transmitted via the 17.7.1 Parasites as food chain that show some degree of specificity for ecological markers their intermediate hosts, thus providing clues about what fish are eating. Finally, parasites can be Parasites can serve as biological tags to obtain in- used as specific characters in phylogenetic recon- formation on various aspects of fish biology (see structions of fish lineages (see Williams et al. 1992) Williams et al. 1992 for an extensive review). Many or in studies of the recent origins of congeneric parasite taxa have all the characteristics of ideal fish species (Marcogliese and Cone 1993). Their biological markers, including infection levels that potential as biological markers is thus vast and vary widely among the different parts of the study mostly untapped. area but remain stable locally from year to year, a relatively long lifespan, and being easily detected 17.7.2 Parasites as early warning and identified (Williams et al. 1992). Parasites systems for pollution detection often provide valuable information on their own, but they should ideally be seen as a comple- Aquatic pollution can influence, directly or indi- mentary method for validating or supplement- rectly, the infection levels and pathogenicity of 384 Chapter 17

fish parasites (see Khan and Thulin 1991 for yet many of these are likely to be of considerable review). On the one hand, pollutants can suppress importance for the way in which we design the immune responses of fish and lead to greater strategies for the conservation, exploitation and acquisition of parasites. On the other, pollutants management of populations. The adoption of a can either harm the free-living stages of fish multidisciplinary approach to fish parasitology, in parasites or reduce the populations of their inver- which controlled experimental infections are used tebrate intermediate hosts and thus cause reduc- to examine changes in the biology of parasitized tions in infection levels in fish. The interaction fish in conjunction with carefully designed studies between toxic pollution and parasitism can there- of host and parasite in more natural environments, fore have important ecological implications for should go some way to filling these gaps. fish populations. Beyond that, fish parasites can In this necessarily brief review, we have tried also be used as indicators in environmental impact to highlight recent research, examining a range of studies and in pollution monitoring. Because topics of current interest, including the role of par- different types of parasites respond differently to asites as agents of natural and sexual selection different pollutants, a detailed knowledge of the in their fish hosts, the consequences of infection- expected effect of pollution on a parasite species associated behaviour change for parasite transmis- is needed if this species is to serve as an indicator. sion, and the effects of parasites on host population For instance, a review of published studies sug- biology. Since many parasites have the potential to gests that ciliate and nematode parasites may be cause direct and indirect losses in commercial and reliable indicators of eutrophication and thermal subsistence fisheries and in aquaculture, and be- pollution, whereas acanthocephalans and cestodes cause others may be transmissible to humans and should make good indicators of heavy metal pol- cause disease or death in extreme cases, we have lution (Lafferty 1997). In fact, acanthocephalans discussed the often considerable socioeconomic and cestodes of fish accumulate heavy metals at importance of infections. We also briefly introduce concentrations that are orders of magnitude higher the use of parasites as indicators of pollution and than those in fish tissues or the external environ- as biological tags and the potential of parasites to ment (Sures et al. 1999). As always, a combination influence fish conservation programmes. of indicators is better than a single one when as- sessing water or environmental quality. The great sensitivity of parasites to many types of pollution REFERENCES should make them indicators of choice. Ackman, R.G. and Gjelstad, R.T. (1975) Gas- chromatographic resolution of isomeric pentanols and 17.8 CONCLUSIONS pentanones in the identification of volatile alcohols and ketones in the codworm Terranova decipiens Fish harbour a wide range of adult and immature (Krabbe, 1878). Analytical Biochemistry 67, 684–7. Arme, C. (1968) Effects of the plerocercoid larvae of forms of taxonomically diverse parasites, many of a pseudophyllidean cestode, Ligula intestinalis, on which have significant sublethal effects that have the pituitary gland and gonads of its host. Biological important consequences for the ecology, evolution Bulletin 134, 15–25. and management of host fish. Parasite infections Arnott, S.A., Barber, I. and Huntingford, F.A. (2000) impact on almost every aspect of fish biology, from Parasite-induced growth enhancement in a fish– individual behaviour to population-level pro- cestode system. Proceedings of the Royal Society of cesses, but there remain many gaps in our know- London Series B 267, 657–63. Austin, B. and Austin, D.A. (1993) Bacterial Fish ledge of the mechanisms by which parasites Pathogens: Diseases of Farmed and Wild Fish. Berlin: influence host individuals and populations. We are Springer-Verlag. only beginning to understand the ecological and Bakke, T.A., Jansen, P.A. and Hansen, L.P. (1990) Differ- fitness consequences of parasites that infect fish, ences in the host resistance of Atlantic salmon, Salmo Fish, Parasites and Disease 385

salar L., stocks to the monogenean Gyrodactylus Bruno, D.W. and Poppe, T.T. (1996) A Colour Atlas of salaris Malmberg, 1957. Journal of Fish Biology 37, Salmonid Diseases. London: Academic Press. 577–87. Butler, J.A. and Millemann, R.E. (1971) Effect of the Ballabeni, P. and Ward, P.I. (1993) Local adaptation of ‘salmon poisoning’ trematode, Nanophyetus salmoni- the trematode Diplostomum phoxini to the European cola, on the swimming ability of juvenile salmonid minnow Phoxinus phoxinus, its second intermediate fishes. Journal of Parasitology 57, 860–5. host. Functional Ecology 7, 84–90. Chandler, M. and Cabana, G. (1991) Sexual dichroma- Balling, T.E. and Pfeiffer, W. (1997) Location-dependent tism in North American freshwater fish: do parasites infection of fish parasites in Lake Constance. Journal play a role? Oikos 60, 322–8. of Fish Biology 51, 1025–32. Chappell, L.H., Hardie, L.J. and Secombes, C.J. (1994) Barber, I. and Crompton, D.W.T. (1997) The distribution Diplostomiasis: the disease and host–parasite inter- of Diplostomum phoxini metacercariae (Trematoda) actions. In: A.W. Pike and J.W. Lewis (eds) Parasitic in the brain of minnows (Phoxinus phoxinus). Folia Diseases of Fish. Dyfed, Wales: Samara Publishing, Parasitologica 44, 19–25. pp. 59–86. Barber, I. and Huntingford, F.A. (1995) The effect of Chubb, J.C. (1976) Seasonal occurrence of helminths in Schistocephalus solidus (Cestoda: Pseudophyllidea) freshwater fishes. Part II: Trematodes. Advances in on the foraging and shoaling behaviour of three-spined Parasitology 17, 111–31. sticklebacks, Gasterosteus aculeatus. Behaviour 132, Coleman, F.C. (1993) Morphological and physiological 1223–40. consequences of parasites encysted in the bulbus arte- Barber, I. and Huntingford, F.A. (1996) Parasite infec- riosus of an estuarine fish, the sheepshead minnow, tion impairs schooling behaviour: spatial deviance of Cyprinodon variegatus. Journal of Parasitology 79, helminth-infected minnows in conspecific groups. 247–54. Proceedings of the Royal Society of London Series B Coombs, I.A. and Crompton, D.W.T. (1991) A Guide to 263, 1095–102. Human Helminths. London: Taylor & Francis. Barber, I. and Ruxton, G.D. (1998) Temporal prey distrib- Costello, M.J. (1996) Development and future of cleaner- ution affects the competitive ability of parasitised fish technology and other biological control tech- sticklebacks. Animal Behaviour 58, 1477–83. niques in fish farming. In: M.D.J. Sayer, J.W. Treasurer Barber, I., Huntingford, F.A. and Crompton, D.W.T. and M.J. Costello (eds) Wrasse: Biology and Use in (1995) The effect of hunger and cestode parasitism on Aquaculture. Oxford: Blackwell Science, pp. 171–84. the shoaling decisions of small freshwater fish. Journal Côté, I.M. (2000) Evolution and ecology of cleaning sym- of Fish Biology 47, 524–36. bioses in the sea. Oceanography and Marine Biology Barber, I., Hoare, D. and Krause, J. (2000) The effects of Annual Review 38, 311–55. parasites on fish behaviour: an evolutionary perspec- Cowell, L.E., Watanabe, W.O., Head, W.D., Grover, J.J. tive and review. Reviews in Fish Biology and Fisheries and Shenker, J.M. (1993) Use of tropical cleaner fish 10, 131–65. to control the ectoparasite Neobenedenia melleni Barber, I., Arnott, S.A., Braithwaite, V.A., Andrew, J. and (Monogenea: Capsalidae) on seawater-cultured Florida Huntingford, F.A. (2001) Indirect fitness consequences red tilapia. Aquaculture 113, 189–200. of mate choice in sticklebacks: offspring of brighter Crompton, D.W.T. and Nickol, B.B. (1985) The Biology males grow slowly but resist parasite infections. of the Acanthocephala. Cambridge: Cambridge Proceedings of the Royal Society of London Series B University Press. 268, 71–6. Crowden, A.E. and Broom, D.M. (1980) Effects of the Bean, C.W. and Winfield, I.J. (1992) Influences of eyefluke, Diplostomum spathaceum, on the behav- the tapeworm Ligula intestinalis (L.) on the spatial iour of dace (Leuciscus leuciscus). Animal Behaviour distributions of juvenile roach Rutilus rutilus (L.) 28, 287–94. and Gudgeon Gobio gobio (L.) in Lough Neagh, Cunningham, E.J., Tierney, J.F. and Huntingford, F.A. Northern Ireland. Netherlands Journal of Zoology 42, (1994) Effects of the cestode Schistocephalus solidus 416–23. on food intake and foraging decisions in the three- Brassard, P., Rau, M.E. and Curtis, M.A. (1982) Parasite- spined stickleback Gasterosteus aculeatus. Ethology induced susceptibility to predation in diplostomiasis. 79, 65–75. Parasitology 85, 495–501. De Meeüs, T., Marin, R. and Renaud, F. (1992) Genetic Bristow, G.A. and Berland, B. (1991) The effect of long- heterogeneity within populations of Lepeophtheirus term, low-level Eubothrium sp (Cestoda, Pseudophyl- europaensis (Copepoda: Caligidae) parasitic on two lidea) infection on growth in farmed salmon (Salmo host species. International Journal for Parasitology 22, salar). Aquaculture 98, 325–30. 1179–81. 386 Chapter 17

Dence, W.A. (1958) Studies on Ligula-infected common Pomphorhynchus laevis. International Journal for shiners (Notropis cornutus frontinalis Agassiz) in the Parasitology 2, 459–69. Adirondacks. Journal of Parasitology 44, 334–8. Hauck, A.K. (1984) Mortality and associated tissues- Dorucu, M., Adams, C.E., Huntingford, F.A. and reactions of chinook salmon, Onchoryhnchus Crompton, D.W.T. (1995) How fish–helminth associa- tshawytscha (Walbaum) caused by the microspridan tions arise: an example from Arctic Charr in Loch Loma sp. Journal of Fish Diseases 7, 217–29. Rannoch. Journal of Fish Biology 47, 1038–43. Hoberg, E.P. (1995) Historical biogeography and modes of El-Matbouli, M. and Hoffmann, R. (1989) Experimental speciation across high latitude seas of the Holarctic: transmission of two Myxobolus spp. developing bis- concepts for host–parasite coevolution among Phocini porogeny via tubificid worms. Parasitology Research (Phocidae) and Tetrabothriidae (Eucestoda). Canadian 75, 461–4. Journal of Zoology 73, 45–57. Evelyn, T.P.T. (1997) A historical review of fish vaccinol- Högland, J. and Thulin, J. (1990) The edidemiology ogy. Developments in Biological Standardization 90, of the metacercariae of Diplostomum baeri and D. 3–12. spathaceum in perch (Perca fluviatilis) from the warm Field, J.S. and Irwin, S.W.B. (1994) The epidemiology, water effluent of a nuclear power station. Journal of treatment and control of diplostomiasis on a fish Helminthology 64, 139–50. farm in Northern Ireland. In: A.W. Pike and J.W. Lewis Holmes, J.C. (1990) Helminth communities in marine (eds) Parasitic Diseases of Fish. Dyfed, Wales: Samara fishes. In: G. Esch, A. Bush and J. Aho (eds) Parasite Publishing, pp. 87–100. Communities: Patterns and Processes. London: Garnick, E. and Margolis, L. (1990) Influence of four Chapman & Hall, pp. 101–30. species of helminth parasites on orientation of Hominick, W. and Chappell, L.H. (1993) The Impact seaward migrating sockeye salmon (Oncorhynchus of Global Change on Disease. Symposia of the nerka) smolts. Canadian Journal of Fisheries and British Society for Parasitology Vol. 30. Cambridge: Aquatic Sciences 47, 2380–9. Cambridge University Press. Giles, N. (1987) Predation risk and reduced foraging Iwama, G. and Nakanishi, T. (1996) The Fish Immune activity in fish: experiments with parasitised and System: Organism, Pathogen and Environment. San non-parasitised three-spined sticklebacks, Gasteros- Diego, CA: Academic Press. teus aculeatus. Journal of Fish Biology 31, 37–44. Jakobsen, P.J., Johnsen, G.H. and Larsson, P. (1988) Effects Goater, C.P. and Holmes, J.C. (1997) Parasite-mediated of predation risk and parasitism on the feeding ecology, natural selection. In: D.H. Clayton and J. Moore habitat use, and abundance of lacustrine threespine (eds) Host–Parasite Evolution: General Principles and stickleback (Gasterosteus aculeatus). Canadian Jour- Avian Models. Oxford: Oxford University Press, pp. nal of Fisheries and Aquatic Sciences 45, 426–31. 9–29. Johnsen, B.O. and Jensen, A.J. (1991) The Gyrodactylus Godin, J.-G.J. and Sproul, C.D. (1988) Risk taking in para- story in Norway. Aquaculture 98, 289–302. sitized sticklebacks under threat of predation: effects Kabata, Z. (1979) Parasitic Copepoda of British Fishes. of energetic need and food availability. Canadian London: Ray Society. Journal of Zoology 66, 2360–7. Kennedy, C.E.J., Endler, J.A., Poynton, S.L. and McMinn, Grutter, A.S. (1999) Cleaner fish really do clean. Nature H. (1987) Parasite load predicts mate choice in guppies. 398, 672–3. Behavioral Ecology and Sociobiology 21, 291–5. Guégan, J.-F. and Agnèse, J.-F. (1991) Parasite evolution- Kennedy, C.R. (1985) Interactions of fish and parasite ary events inferred from host phylogeny: the case of populations: to perpetuate or pioneer? In: D. Rollinson Labeo species (Teleostei, Cyprinidae) and their dactyl- and R.M. Anderson (eds) Ecology and Genetics of ogyrid parasites (Monogenea, Dactylogyridae). Cana- Host–Parasite Interactions. London: Academic Press, dian Journal of Zoology 69, 595–603. pp. 1–20. Guthrie, J.F. and Kroger, R.L. (1974) Schooling habits Kennedy, C.R. (1994) The ecology of introductions. In: of injured and parasitised menhaden. Ecology 55, 208– A.W. Pike and J.W. Lewis (eds) Parasitic Diseases of 10. Fish. Dyfed, Wales: Samara Publishing, pp. 189–208. Hamilton, W.D. and Zuk, M. (1982) Heritable true fitness Kennedy, C.R. (1995) Richness and species diversity of and bright birds: a role for parasites? Science 218, macroparasite communities in tropical eels Anguilla 384–7. reinhardtii in Queensland, Australia. Parasitology Hardisty, M.W. and Potter, I.C. (1971) The Biology of the 111, 233–45. lampreys. London: Academic Press. Kennedy, C.R., Wyatt, R.J. and Starr, K. (1994) The de- Harris, J.E. (1972) The immune response of a cline and natural recovery of an unmanaged coarse cyprinid fish to infections of the acanthocephalan fishery in relation to changes in land use and attendant Fish, Parasites and Disease 387

eutrophication. In: I.G. Cowx (ed.) Rehabilitation of Matthews, R.A. (1994) Ichthyophthirius multifilis Freshwater Fisheries. Oxford: Fishing News Books, pp. Fouquet, 1876: infection and protective response 366–75. within the fish host. In: A.W. Pike and J.W. Lewis Kent, M.L., Andree, K.B., Bartholomew, J.L. et al. (2001) (eds) Parasitic Diseases of Fish. Dyfed, Wales: Samara Recent advances in our knowledge of the Myxozoa. Publishing, pp. 17–42. Journal of Eukaryotic Microbiology 48, 395–413. Milinski, M. (1984) Parasites determine a predator’s Khalil, L.F., Jones, A. and Bray, R.A. (1994) Keys to optimal feeding strategy. Behavioral Ecology and the Cestode Parasites of Vertebrates. London: CAB Sociobiology 15, 35–7. International. Milinski, M. (1985) Risk of parasitised sticklebacks Khan, R.A. and Thulin, J. (1991) Influence of pollution on (Gasterosteus aculeatus L.) under competition for parasites of aquatic animals. Advances in Parasitology food. Behaviour 93, 203–15. 30, 201–38. Milinski, M. (1990) Parasites and host decision-making. Kodric-Brown, A. (1990) Mechanisms of sexual selection: In: C.J. Barnard and J.M. Behnke (eds) Parasitism insights from fishes. Annales Zoologi Fennici 27, and Host Behaviour. London: Taylor and Francis, pp. 87–100. 95–116. Krause, J. and Godin, J.-G.J. (1994) Influence of parasitism Moles, A. and Heifetz, J. (1998) Effects of the brain on the shoaling behaviour of banded killifish, Fun- parasite Myxobolus arcticus on sockeye salmon. dulus diaphanus. Canadian Journal of Zoology 72, Journal of Fish Biology 52, 146–51. 1775–9. Moody, J. and Gaten, E. (1982) The population dynamics Lafferty, K.D. (1997) Environmental parasitology: what of eyeflukes Diplostomum spathaceum and Tylodel- can parasites tell us about human impacts on the envi- phys clavata (Digenea: Diplostomatidae) in rainbow ronment? Parasitology Today 13, 251–5. and brown trout in Rutland Water: 1974–1978. Lafferty, K.D. (1999) The evolution of trophic transmis- Hydrobiologia 88, 207–9. sion. Parasitology Today 15, 111–15. Moravec, F. (1992) Spreading of the nematode Anguil- Lafferty, K.D. and Morris, A.K. (1996) Altered behavior licola crassus (Dracunculoidea) among eel popu- of parasitized killifish increases susceptibility to lations in Europe. Folia Parasitologica 39, 247–8. predation by bird final hosts. Ecology 77, 1390–7. Moravec, F. (1994) Parasitio nematones of freshwater LoBue, C.P. and Bell, M.A. (1993) Phenotypic manipula- fishes of Europe. Dordrecht: Kluwer Academic tion by the cestode parasite Schistocephalus solidus Publishers. of its intermediate host, Gasterosteus aculeatus, Moser, M. (1991) Parasites as biological tags. Parasitol- the threespine stickleback. American Naturalist 142, ogy Today 7, 182–5. 725–35. Moser, M. and Kent, M.L. (1994) Myxosporea. In: Lom, J. and Dyková, I. (1992) Protozoan Parasites of J.P. Kreier (ed.) Parasitic Protozoa, 2nd edn, Vol. 8, Fishes. Amsterdam: Elsevier. New York: Academic Press, pp. 265–318. López, S. (1999) Parasitized female guppies do not prefer Ness, J.H. and Foster, S.A. (1999) Parasite-associated showy males. Animal Behaviour 57, 1129–34. phenotype modifications in threespine stickleback. Losey, G.S. (1987) Cleaning symbiosis. Symbiosis 4, Oikos 85, 127–34. 229–58. Owen, S.F., Barber, I. and Hart, P.J.B. (1993) Low level Lysne, D.A., Hemmingsen, W. and Skorping, A. (1997) infection by eyefluke, Diplostomum spp. affects Regulation of infrapopulations of Cryptocotyle lingua the vision of three-spined sticklebacks, Gasterosteus on cod. Parasitology 114, 145–50. aculeatus. Journal of Fish Biology 42, 803–6. Madhavi, R. and Anderson, R.M. (1985) Variability in Pampoulie, C., Morand, S., Lambert, A., Rosecchi, E., the susceptibility of the fish host, Poecilia reti- Bouchereau, J.L. and Crivelli, A.J. (1999) Influence culata, to infection with Gyrodactylus bullatarudis of the trematode Aphalloïdes coelomicola Dollfus, (Monogenea). Parasitology 91, 531–44. Chabaud & Golvan, 1957 on the fecundity and survival Marcogliese, D.J. and Cone, D.K. (1993) What metazoan of Pomatoschistus microps (Krøyer 1838) (Teleostei: parasites tell us about the evolution of American and Gobiidae). Parasitology 119, 61–7. European eels. Evolution 47, 1632–5. Paterson, A.M. and Poulin, R. (1999) Have chondracan- Marcogliese, D.J. and Cone, D.K. (1997) Food webs: a thid copepods co-speciated with their teleost hosts? plea for parasites. Trends in Ecology and Evolution Systematic Parasitology 44, 79–85. 12, 320–5. Poulin, R. (1992) Determinants of host specificity Markiw, M.E. (1992) Salmonid whirling disease. US in parasites of freshwater fishes. International Journal Fisheries and Wildlife Service Leaflet 17, Washington, for Parasitology 22, 753–8. DC. Poulin, R. (1993) Age-dependent effects of parasites on 388 Chapter 17

anti-predator responses in two New Zealand freshwa- Roberts, R.J. (1989) Fish Pathology, 2nd edn. London: ter fish. Oecologia 96, 431–8. Baillière Tindall. Poulin, R. (1994) Mate choice decisions by parasitized Rodewald, A.D. and Foster, S.A. (1998) Effects of gra- female upland bullies Gobiomorphus breviceps. Pro- vidity on habitat use and anti-predator behaviour in ceedings of the Royal Society of London Series B three-spined sticklebacks. Journal of Fish Biology 52, 256, 183–7. 973–84. Poulin, R. (1995) ‘Adaptive’ changes in the behaviour of Rohde, K. (1993) Ecology of Marine Parasites, 2nd edn. parasitized animals: a critical review. International Wallingford: CAB International. Journal for Parasitology 25, 1371–83. Rosenqvist, G. and Johansson, K. (1995) Male avoidance Poulin, R. (1998) Evolutionary Ecology of Parasites: of parasitized females explained by direct benefits in a From Individuals to Communities. London: Chapman pipefish. Animal Behaviour 49, 1039–45. & Hall. Sakanari, J.A. and Moser, M. (1990) Adaptation of an Poulin, R. and FitzGerald, G.J. (1989a) Risk of parasitism introduced host to an indigenous parasite. Journal of and microhabitat selection in juvenile sticklebacks. Parasitology 76, 420–3. Canadian Journal of Zoology 67, 14–18. Sawyer, R.T. (1986) Leech Biology and Behaviour. Poulin, R. and FitzGerald, G.J. (1989b) Shoaling as Oxford: Clarendon Press. an anti-ectoparasite mechanism in juvenile stickle- Schmidt, G.D. and Roberts, L.S. (1989) Foundations backs (Gasterosteus spp.). Behavioral Ecology and of Parasitology. St Louis: Times Mirror/Mosby Sociobiology 24, 251–5. College Publishing. Poulin, R. and Grutter, A.S. (1996) Cleaning symbioses: Sharp, G.J.E., Pike, A.W. and Secombes, C.J. (1989) The proximate and adaptive explanations. BioScience 46, immune response of wild rainbow trout Salmo gaird- 512–17. neri Richardson, to naturally acquired plerocercoid Poulin, R. and Rohde, K. (1997) Comparing the richness infections of Diphyllobothrium dendriticum of metazoan ectoparasite communities of marine (Nitzsch, 1824) and D. ditremum (Creplin, 1825). fishes: controlling for host phylogeny. Oecologia 110, Journal of Fish Biology 35, 781–94. 278–83. Sikkel, P.C. (1986) Intraspecific cleaning by juvenile Radabaugh, D.C. (1980) Changes in minnow, Pimephales salema, Xenestius californiensis (Pisces: Haemulidae). promelas Rafinesque, schooling behaviour associated California Fish and Game 72, 170–2. with infections of brain-encysted larvae of the fluke, Sinderman, C.J. (1990) Principal Diseases of Marine Ornithodiplostomum ptychocheilus. Journal of Fish Fish and Shellfish, 2nd edn. San Diego, CA: Academic Biology 16, 621–8. Press. Rahimian, H. and Thulin, J. (1996) Epizootiology of Smith, H.D. (1973) Observations on the cestode Euboth- Ichthyophonus hoferi in herring populations off the rium salvelini in juvenile sockeye salmon (On- Swedish west coast. Disease of Aquatic Organisms 27, corhynchus nerka) at Babine Lake, British Columbia. 187–95. Journal of the Fisheries Research Board of Canada 30, Ranta, E. (1995) Schistocephalus infestation improves 947–64. prey-size selection by three-spined sticklebacks, Smith, H.D. and Margolis, L. (1970) Some effects Gasterosteus aculeatus. Journal of Fish Biology 46, of Eubothrium salvelini (Schrank, 1790) on sockeye 156–8. salmon, Oncorhynchus nerka (Walbaum), in Babine Rees, G. (1955) The adult and Diplostomulum stage Lake, British Columbia. Journal of Parasitology 56, (Diplostomulum phoxini (Faust)) of Diplostomum 321–2. pelmatoides Dubois and an experimental dem Sommerville, C. (1998) Parasites of farmed fish. In: onstration of part of the life cycle. Parasitology 45, K.D. Black and A.D. Pickering (eds) Biology of Farmed 295–311. Fish. Sheffield: Sheffield Academic Press, pp. 146– Renaud, F. and Gabrion, C. (1988) Speciation in Cestoda: 79. evidence of two sibling species in the complex Sprengel, G. and Lüchtenberg, H. (1991) Infection by en- Bothrimonus nylandicus (Scheider, 1902) (Cestoda, doparasites reduces maximum swimming speed of Pseudophyllidea). Parasitology 97, 1–9. European smelt Osmerus eperlanus and European eel Rigby, M.C. Holmes, J.C., Cribb, T.H. and Morand, S. Anguilla anguilla. Diseases of Aquatic Organisms 11, (1997) Patterns of species diversity in the gastro- 31–5. intestinal helminths of a coral reef fish, Epinephelus Sproston, N.G. and Hartely, P.H.T. (1941) The ecology of merra (Serranidae), from French Polynesia and the some parasitic copepods of gadoids and other fishes. South Pacific Ocean. Canadian Journal of Zoology 75, Journal of the Marine Biological Association of the UK 1818–27. 25, 361–92. Fish, Parasites and Disease 389

St-Hilaire, S., Hill, M., Kent, M.L., Whitaker, D.J. and Fish vaccines. Advances in Veterinary Medicine 41, Ribble, C. (1997) A comparative study of muscle 539–50. texture and intensity of Kudoa thyrsites infection in Ward, P.I. (1988) Sexual dichromatism and parasitism in farm-reared Atlantic salmon Salmo salar on the British and Irish freshwater fish. Animal Behaviour 36, Pacific coast of Canada. Disease of Aquatic Organisms 1210–15. 31, 221–5. Wedekind, C. and Milinski, M. (1996) Do three-spined Stott, M.K. and Poulin, R. (1996) Parasites and parental sticklebacks avoid consuming copepods, the first care in male upland bullies (Eleotridae). Journal of Fish intermediate host of Schistocephalus solidus? An Biology 48, 283–91. experimental analysis of behavioural resistance. Sures, B., Siddall, R. and Taraschewski, H. (1999) Para- Parasitology 112, 371–83. sites as accumulation indicators of heavy metal pollu- Williams, H.H. and Jones, A. (1994) Parasitic Worms of tion. Parasitology Today 15, 16–21. Fish. London: Taylor & Francis. Sweeting, R.A. (1977) Studies on Ligula intestinalis: Williams, H.H., MacKenzie, K. and McCarthy, A.M. some aspects of the pathology in the second interme- (1992) Parasites as biological indicators of the popula- diate host. Journal of Fish Biology 10, 43–50. tion biology, migrations, diet, and phylogenetics of Szidat, L. (1969) Structure, development and behaviour fish. Reviews in Fish Biology and Fisheries 2, 144–76. of new strigeatoid metacercariae from subtropical Williams, M.A., Penlington, M.C., King, J.A., Hoole, D. fishes of South America. Journal of the Fisheries and Arme, C. (1998) Ligula intestinalis (Cestoda) in- Research Board of Canada 26, 753–86. fections of roach (Rutilus rutilus) (Cyprinidae): im- Tierney, J.F. (1994) Effects of Schistocephalus solidus munocytochemical investigations into the salmon- (Cestoda) on the food intake and diet of the three- and chicken-II type gonadotrophin-releasing hormone spined stickleback, Gasterosteus aculeatus. Journal of (GnRH) systems in host brains. Acta Parasitologica Fish Biology 44, 731–5. 43, 232–5. Tierney, J.F., Huntingford, F.A. and Crompton, D.W.T. Wolf, K. (1988) Fish Viruses and Viral Diseases. Ithaca, (1993) The relationship between infectivity of NY: Cornell University Press. Schistocephalus solidus (Cestoda) and anti-predator Woo, P.T.K. (1995) Fish Diseases and Disorders. Vol. 1: behaviour of its intermediate host, the three- Protozoan and Metazoan Infections. London: CAB spined stickleback, Gasterosteus aculeatus. Animal International. Behaviour 46, 603–5. Woo, P.T.K. and Bruno, D.W. (1999) Fish Diseases and van Dobben, W.H. (1952) The food of the cormorant in Disorders. Vol. 3: Viral, Bacterial and Fungal Infec- the Netherlands. Ardea 40, 1–63. tions. London: CAB International. Videler, J.J. (1993) Fish Swimming. London: Chapman Wyman, R.L. and Walters-Wyman, M.F. (1985) Chafing & Hall. in fishes: occurrence, ontogeny, function and evolu- Vinitnantharat, S., Gravningen, K. and Greger, E. (1999) tion. Environmental Biology of Fishes 12, 281–9. Index

Note: Numbers in italics indicate figures; numbers in bold indicate tables. Cross-references to Volume 2 appear at the end of the relevant entry under see also, and give both volume number and page number. abundance surveys Adrianichthyidae (ricefishes), 34 alarm substances, 18, 28, 286–7 density-dependent mortality Adrianichthyinae, 54 albacore tuna (Thunnus alalunga), estimation, 133, 135–6 Aequidens curviceps, 256 stock structure genetic key-factor analysis limitations, Aequidens pulcher (blue acara), 287 analysis, 214 133, 135 aerobic metabolism Alburnus alburnus (bleak), 287 marine larval stage sampling critical temperature thresholds, Albutiformes, 19, 22, 28 problems, 138 87 alepisaurids (lancetfishes), 30 migration as source of error, 194 swimming, 77 alewife (Alosa pseudoharengus), 136 see also 2.130, 2.141, 2.145–6, aestivation, 90 algae, 351 2.166, 2.172, 2.243 Africa, Ethiopian region freshwater see also phytoplankton Abyssocottidae, biogeography, 50 biogeography, 51–2 alimentary tract, 90 acanthocephalans, 360, 361 African lungfishes see carnivorous fish, 90–1 Acanthomorpha, 30, 31 Lepidosireniformes herbivorous fish, 90, 91 Acanthuridae, biogeography, 59 Agassiz, Louis, 20 Allee effect, 6, 7, 9, 2.331–2 Acanthuroidei, 35 age at maturity, 6, 149, 151, 152–3 see also depensation Acanthurus (surgeonfish), 91 growth rate relationship, 103, 157, alloparents, 237 acid–base regulation, 83 166 allozyme analysis, 200, 202 freshwater fish, 78, 79 life history invariants, 158 brown trout (Salmo trutta) Acipenser oxyrhynchus (Atlantic piscivore life history strategies, introgression following stock sturgeon), 158 270 enhancement programmes, 216 Acipenser spp., sex determination, response to fishing, 165, 166, 167 fish tissue identification, 206, 207 218 age determination, 104–7, 2.72 genetic diversity studies, 208–9; Acipenser sturio (common annual marks/rings in hard parts, connectivity between sturgeon), 99 105–6 subpopulations, 208, 208; caviar identification: DNA back calculation, 107 fishing-related reduction, 215 analysis, 207; isoelectric direct observations, 104–5 hybridization studies, 207 focusing, 206 elasmobranchs, 106–7 stock structure: chum salmon see also 2.326 length–frequency analysis, 105 (Oncorhynchus keta), 211, 212, Acipenseridae see also 2.188–208 212; cod (Gadus morhua), 212, anadromy, 187 age-related parasite burden, 375 213; yellowfin tuna (Thunnus distribution, 193 age-specific fecundity, 149, 150 albacares), 214 see also 2.198, 2.330, 2.331 age-specific survival, 149, 150 allozyme heterozygosity (h) per Acipenseriformes, 19, 22, 27 Agnatha, 71 locus, 208, 208 biogeography, 50 agonids, biogeography, 58 Alopidae (thresher sharks), 24–5 acoustic surveys, 194 Aidablennius sphynx, 233 Alosa pseudoharengus (alewife), 136 Actinopterygii, 26 air-breathing organs, 89–90 Alosa sapidissima (American shad), acylglycerol, static lift contribution, Alabes, 35 113, 152–3 73 alarm response, 18, 28, 286 Alosa spp., 136 Index 391 alternative reproductive strategies, anemonefish (Amphiprion biological control, 382–3; 162–5 akallopisos), 239 microparasites, 377 adaptive phenotypic plasticity, angelsharks see Squatiniformes see also 2.38, 2.39, 2.40, 2.41, 2.46, 164 anglerfish (Lophius spp.), 251 2.49, 2.53, 2.54, 2.55, 2.56, definition, 162 anglerfishes see Lophiiformes 2.324–5 fitness impact evaluation, 162–3 Anguilla anguilla (European eel), 75, aquatic adaptations, 71–93 genetic basis, 163 87, 153, 180 bioluminescence, 91–2 maintenance, 163–5; growth-rate migratory behaviour, 178, 180, buoyancy, 71, 72–5 thresholds, 164–5; negative 190 digestive function, 90–1 frequency-dependent selection, Anguilla dieffenbachii (longfin eel), locomotion, 71, 75–7 163, 163 188 low temperature/freezing, 81–2 management implications, 165 Anguilla japonica (Japanese eel), 180 osmoregulation, 78–81 Amarillo fish (Girardinichthys Anguilla spp. (eel), 98, 99 oxygen availability, 82–3 multiradiatus), 227, 233 cutaneous gas exchange, 84 respiratory function, 82–90 Amazon basin, 53 glass-eel commercial Arapaima gigas (arapaima; Ambloplites rupestris (rock bass), expolitation, 179 pirarucú), 27, 53 287, 289 migratory behaviour, 78, 186, 190; Arctic char (Salvelinus alpinus), Amblyopsidae (cavefishes), 32 ocean currents influence, 156, 158, 234 Amblyopsiformes, 32 180–1; scale of movement, 176 brown trout competitive ambush tactics, 268, 271 parasite latitudinal variation, 363 interactions, 325 American knifefishes see Anguillicola carssus, 376 functional responses, 273 Gymnotiformes Anguillidae (freshwater eels) light-related feeding behaviour, American shad (Alosa sapidissima), catadromy, 188, 193 324 113, 152–3 life cycle, 188 temperature-related performance, Amia calva, 50 anguilliform swimming, 75 324 Amiidae, biogeography, 50, 51 Anguilliformes, 19, 22, 28 trophic morphs, 270, 365 Amiiformes, 19, 22, 27 anisakiasis, 377 area cladograms, 47, 48, 50 amphidromy, 188, 189 anisakid human pathogens, 377 construction methods, 47 definition, 177, 178 Anisakis simplex, 9, 362–3 Argentiniformes, 19, 22, 29, 30 exploitation of migration pattern, Anisakis sp., 377 Ariidae (forktailed catfishes) 178 annual marks/rings (annuli), 105–6 anadromy, 187 amphioxus, 75 annual reproductive rate, 6, 130 biogeography, 55, 58 Amphiprion akallopisos Anomalopidae (flashlight fishes), 33, armoured catfish (Loricaridae), 257, (anemonefish), 239 92 332 amplified fragment length Anoplopomatidae, biogeography, Arripidae, biogeography, 61 polymorphism (AFLP), 205 58 Artedidraconidae, biogeography, 61 amylase, 90 Antarctic Ocean artificial spawning sites, 240–1 Anabantoidei, 35 biogeography, 61–2 asexual reproduction, 218, 237–8 Anacanthini, 32 physiological adaptations, 61–2, Asian Pacific region, 344 anadromy, 187–8, 189 82; isohaemoglobins, 87 Ateleopodiformes, 19, 22, 30 definition, 176, 177 antennariids (frogfishes), 33 Atherina spp., 287 genetic diversity studies, 208, 208, biogeography, 61 Atherinidae 209 luring behaviour, 268 biogeography, 55, 56, 61 homing behaviour, 190, 191 antibodies, 382 environmental sex larval stage density-dependent ‘antifreeze’ compounds, 81–2 determination, 218 mortality, 136, 138 antioxidative defences, 91 grouping as prey defence, 287 spawning aggregate Apache trout (Oncorhynchus Atheriniformes, 19, 23, 31, 34 overexploitation, 240 apache), 216 Atherinomorpha, 31, 34 stable isotope studies, 306 Aphredoderidae (pirate perches), 32 Atherinopsidae, biogeography, 50 stock structure, 211 Aphredoderoidei, 32 Atlantic cod see Gadus morhua see also 2.349 Apistogramma, 218 Atlantic coral reef (Eastern anaerobic metabolism aplocheiloids (killifish), Atlantic), 344 critical temperature thresholds, biogeography, 52 Atlantic herring see Clupea 87 Aplochitonidae, 187, 188 harengus swimming, 77 Aplodactylidae, biogeography, 61 Atlantic salmon see Salmo salar Ancharias, 52 Apogonidae, biogeography, 58, 59, Atlantic silverside (Menidia Anchoa mitchilli (bay anchovy), 61 menidia), 113, 155 274 apomorphic characters, 16 Atlantic sturgeon (Acipenser anchor tag, 105 aquaculture, 2, 3 oxyrhynchus), 158 anchovies see Clupeiformes parasitism impact, 376–7; Aulopiformes, 19, 22, 30 392 Index aulorhynchids, biogeography, 58 chasing behaviour, 268 bluehead wrasse (Thalassoma Australia DNA analysis for species bifasciatum), 164, 227, 241 freshwater biogeography, 54–5 identification, 207 protogynous hermaphroditism, southern marine fauna, 61 migration, 176 218 autoapomorphy, 16 piscivory, 267 body condition autosomal sex determination, 218 biodiversity, 4–5, 19, 47, 49 parental care fitness costs, 228 Auxis rochei (bullet mackerel), 72 diadromous migratory species, refuge use, 291 ayu (Plecoglossus altivelis), 178 193–4 body depth amphidromy, 188 fish parasite communities, 363–4 morphological prey defences, global clines, 343 286 bacteria, 360, 361 hotspots, 343 piscivorous fish, 269 Bagridae, biogeography, 54 littoral fish faunas: temperate, body moisture, tissue lipids Balistidae (triggerfish), 251 344–5; tropical, 343–4 relationship, 108, 109 see also 2.306, 2.344 marine species, 342 body size, 8, 97 Balitoridae, biogeography, 54 trends with latitude, 342–3, 343 at maturity, 149, 151, 152–3; banded killifish (Fundulus biogeographic realms, 44, 45 influence of fishing gear, 167; diaphanus), 288 biogeography, 5, 44, 342 life history invariants, 158; barbels, DNA analysis for species coral reef, 344 response to fishing, 166 identification, 207 migratory fish, 189, 192–4 competition, 8, 326–7, 329; barramundi (Lates calcarifer) temperate regions, 342 intercohort, 334–5, 335 catadromy, 188 tropical regions, 342 growth curves, 103 protandrous hermaphroditism, see also historical biogeography length data see length data 218 biological tagging, 383 metabolic rate relationship, 334 barred serrano (Serranus fasciatus), bioluminescence, 91–2 ontogenetic niche shifts, 327–8 236, 237 self-luminescence, 92 parasite burden relationship, 365, basking shark (Cetorhinus symbiotic luminous bacteria, 92 375 maximus), 24 biomanipulation, piscivore population structure, 321–2; Batesian mimicry, 285 introductions for water quality productivity, 328–30, 329 Bathydraconidae, biogeography, 61 improvement, 279 predator–prey interactions, Batoidea see Rajiformes bitterling see European bitterling 275–7, 276, 327, 335–6 Batrachoidiformes, 19, 22, 31, 32, 33 black marlin (Makaira indica), 35 prey, 258, 274, 275, 286; predator bay anchovy (Anchoa mitchilli), 274 black-chinned tilapia (Sarotherodon size relationships, 275–7, 276; beardfishes see Polymixiiformes melanotheron), 235 refuge use, 292, 292 Bedotia, 52 blacknose dace (Rhinicthys sexual selection, 230–1 Bedotiidae, 34 atratulus), 258 see also 2.294, 2.296, 2.297, 2.303, biogeography, 52 bleak (Alburnus alburnus), 287 2.342–3 behavioural characteristics Blenniidae (blennies) Bohr effect, 85, 87 parasitism-related changes, Batesian mimicry, 285 bomb calorimetry, 111–12 367–74, 370–3 biogeography, 56, 59 bonefish see Albutiformes systematic investigations, 17–18 migratory movements, 176 bonito (Sarda sarda), 72 see also 2.230, 2.233–6 Blennioidei, 35 bony fishes (Teleostomi), 21, 26 Belonidae (needlefishes), 34 biogeography, 58 bonytongues see Osteoglossiformes stalking behaviour, 268 blood osmolality, 78 bottlenecks, 215 Beloniformes, 19, 23, 31, 34 blood-sucking habit, 268 Bovichthyidae (Bovichtidae) Bermuda, 57, 58 blue acara (Aequidens pulcher), 287 biogeography, 61 Beryciformes, 19, 23, 31, 33 blue shark (Prionace glauca), 77, 189 catadromy, 188 Beverton–Holt stock-recruitment blue whiting (Micromesistius bower-building cichlid (Cyrtocara model, 125, 128, 129 poutassou), 106 eusinostomus), 236, 240 see also 2.75, 2.76, 2.118, 2.127, see also 2.19 bowfins see Amiiformes 2.178, 2.179, 2.204, 2.272, blue/fin whale hybrid (Balaenoptera boxfishes, 75 2.283, 2.287 musculus x B. physalus), 207 brachionichthyids (handfishes), 33 bichirs (Polypteridae), biogeography, bluefin tuna (Thunnus thynnus) Brachydanio rerio (zebra danio; 51 migratory behaviour, 182–3, 189 zebrafish), 29 bicolor damselfish (Stegastes see also 2.21, 2.23, 2.204, 2.330 Brachyplatystoma, 53 partitus), 233 bluefish (Pomatomus saltatrix), 76, bramble sharks see bigeye tuna (Thunnus obesus) 268–9 Echinorhiniformes stock structure genetic analysis, prey selection, 274 branchiuran parasites, 360, 361 214, 215 bluegill sunfish (Lepomis branding, 191 see also 2.21, 2.22, 2.23 macrochirus), 8, 153, 255, 255, breeding systems, 225–7, 226 billfishes, 342 257, 258, 325, 326, 327, 330, 332 environmental influences, 227 Index 393

reproductive success carrying capacity, 130–1 Chilean jack mackerel (Trachurus determinants, 226 spawner–recruit relationship, 124 murphyi), 2 see also mate choice; mating variability, 130, 131 Chimaeriformes, 19, 21, 21, 24 patterns; sexual selection see also 2.105, 2.107, 2.109, 2. chinook salmon (Oncorhynchus Brody equation, 103 110, 2.111, 2.114, 2.118, 2.119, tshawytscha), 167, 187, 193 brooding behaviour, 228 2.121 see also 2.379 brook char see Salvelinus fontinalis cartilaginous fishes see Chironemidae, biogeography, 61 brook trout see Salvelinus fontinalis Chondrichthyes chloride cells, 81, 82 brown trout see Salmo trutta catadromy, 78, 188, 189 Chondrichthyes (cartilaginous bulbus, 84 definition, 176, 177, 178 fishes), 21, 71 bullet mackerel (Auxis rochei), 72 juvenile stages exploitation, 179 egg size, 151 buoyancy, 71, 72–5 Catastomidae, biogeography, 50 offspring size/numbers, 158 dynamic lift, 72 catfishes (Siluriformes), 19, 22, 29 Chrysiptera cyanea, 231 static lift, 72–5; swimbladders, biogeography, 53, 54 Chrysophyceae, fatty acid analysis, 73–5; tissue lipids, 73 egg size, 151 309 butterflyfish (Pantodon), 51 parasitism (Trichomycteridae and chub mackerel, 76 Cetopsidae), 268 chum salmon (Oncorhynchus keta), 13C isotopes, food web analysis, 306 catsharks (Scyliorhinidae), 24 187 Caligus spp., 362 caudal pump (caudal hearts), 84 stock structure genetic analysis, Callionymidae, biogeography, 61 cavefishes, 19, 22, 31, 32 211–12, 212 Calloplesiops altivelis, 285 caviar identification Cichlasoma citrinellum (Midas candirús (parasitic catfishes), 53 DNA analysis, 207 cichlid), 237 canned fish, 2 isoelectric focusing, 206 Cichlasoma nigrofasciatum cannibalism, 278, 334, 336, 336–7 cellulose-acetate protein (convict cichlid), 227, 237 see also 2.239, 2.345, 2.378 fingerprints, 206 Cichlidae (cichlids) capelin (Mallotus villosus), 160 Central America, freshwater alternative reproductive see also 2.16, 2.19, 2.184, 2.236, biogeography, 52–4 strategies, 162 2.241, 2.242–3 Centrarchidae, 329 biogeography, 50, 51, 52, 53 capture fisheries production, 1–2, 2 biogeography, 50 communal breeding, 237 capture success, 254, 271–2, 272 Centriscidae (shrimpfishes), 35 East African lake species flocks, body size relationship, 275 Centropomidae (snooks), 51–2 grouping as prey defence, 287 catadromy, 188 extrabodily ornaments, 234 temperature influence, 324, 324 Cephalorhynchus hectori (Hector’s isohaemoglobins, 87 Carangidae (jacks), 75 dolphin), 215 molecular philogenetic studies, biogeography, 56, 58 Ceratioidei, 33, 92 52 see also 2.304, 2.343 Ceratodontiformes, 19, 22, 26 parental care, 230 carangiform swimming, 75 cestode parasites, 360, 361, 362, 363, see also 2.325 Carassius auratus (goldfish), 29, 256 368, 377 circulatory system, 84 biogeography, 50 human pathogens, 377, 378–81 Cirripectes, 56 Carassius carassius (crucian carp), cetacean predation, 287 cisco (Coregonus albula; vendace), 286 Cetorhinus maximus (basking 335, 336 carbon dioxide solubility, 82–3 shark), 24 Citharinidae, biogeography, 51 Carcharhiniformes, 19, 21, 24 Chaenopsidae, biogeography, 58 clades, 17 Carcharhinus (whaler/requiem Chaetodontidae, biogeography, 59 Cladistia see Polypteriformes shark), 24 Chagos Archipelago, 60 cladistics, 15, 16 Carcharodon carcharias (great Chandidae, biogeography, 55 biogeographic applications, 44, white shark), 24 channel catfish, sex determination, 46, 47; area cladograms, 47, 48 cardinalfish, Batesian mimicry, 285 218 character-state homologies, 16–17 cardiovascular adaptations for Channichthyidae, biogeography, 61, clades (monophyletic taxa), 17 swimming, 77 87 Clarias, 218 Caribbean Characidae, biogeography, 51, 52 Clariidae, biogeography, 51 biogeography, 57, 58 Characiformes, 19, 22, 29 Clarioteidae, biogeography, 51 coral reef, 344 biogeography, 53, 54 class, taxonomic, 18 carnivorous fishes characins see Characiformes classification alimentary tract structure, 90–1 Charles’s categories of uncertainty, cladistic, 16–17 energy budget, 110 3 extant fish orders, 19, 21–3 see also piscivorous fishes see also 2.4, 2.5 hierarchical ranking system, 18, carp see Cyprinus carpio chasing behaviour, 268 18–20; intraspecific ranks, 20 carpet sharks see Orectolobiformes chemical markers, 105, 191 methods, 15 carps see Cypriniformes chemical parasite control, 383 phenetic, 15, 16 394 Index cleaning symbiosis, 382 Comephoridae, biogeography, 50 cooperative hunting behaviour, 269 climate change, 5 common goby (Pomatoschistus copepod parasites, 360, 361, 362 parasite infection impact, 376 microps), 156, 228 coral reef sea bass population response common sturgeon (Acipenser biodiversity, 343, 344, 346; study prediction, 185 sturio), 99 methods, 345 climbing, upstream migration, 193 caviar identification: DNA biogeography, 344 clingfishes see Gobiesociformes analysis, 207; isoelectric demersal species, 346–7 Clinidae, biogeography, 61 focusing, 206 food webs, 313–14, 314; stable closed areas, 194 communal breeding, 237 isotope studies, 306, 307, 308 see also 2.14, 2.379 community ecology, 8–9, 321–37 management approaches, 347 Clupea harengus (Atlantic herring), abiotic factors, 323–5 recruitment, 123, 351; 8, 9, 73, 76, 190, 200, 275 competitive interactions, 325–7; spatiotemporal variation, 351–2 egg size, 160 size-dependent, 326–7, 329, see also 2.19, 2.212, 2.294, 2.296, food web, 303 334–5, 335 2.298, 2.300, 2.304, 2.307, migratory behaviour, 176, 180, habitat structure, 330–1 2.309, 2.310, 2.327–8, 2.329, 190 large-scale patterns, 322–3 2.331, 2.343, 2.344, 2.347, 2.354 parasites, 9 marine commmunities, 341–53 coral trout (Plectropomus population recovery from low productivity, 328–30; size- areolatus), 241 spawner abundance, 125 structured processes, 329 Coregonidae, biogeography, 49 spawning aggregates, 236; size-dependent mixed Coregonus albula (cisco; vendace), overexploitation, 240 interactions, 327 335, 336 see also 2.16, 2.19, 2.49, 2.55, 2.72, see also 2.306, 2.342, 2.343, corkwing wrasse (Symphodus 2.73, 2.156, 2.160, 2.176, 2.242, 2.344–5, 2.352–3 melops), 163 2.251, 2.330, 2.345, 2.347, 2.348 compensatory reserve, 6, 124 Coryphaenoides rupestris (round- Clupea harengus pallasi, 278 experimental study approaches, nosed grenadier), 194 Clupeidae 141 Corythoichthys intestinalis amphidromy, 188 meta-analysis, 129–30 (pipefish), 237 anadromy, 187 role of specific life history stages, Costa Rica, 53 biogeography, 49 133 Cottidae (sculpins) catadromy, 188 stock-recruitment models, 124–5 amphidromy, 188 spawner–recruit relationship, 125 see also maximum reproductive biogeography, 50, 58, 344 Clupeiformes, 19, 22, 28 rate catadromy, 188 anadromy, 187 competition, 325–7 home range, 349 Clupeocephala, 28 age/size-related changes, 8 Cottus cognatus (slimy sculpin), 227 Clupeomorpha, 28, 73 asymmetry, 328 Cottus gobio (river bullhead), 228, coalfish, 76 habitat structure effects, 330, 331 234 coastal zone productivity, 303–4, individual-based models, 142 counter-shading, 285 304 mating, 225, 226–7, 230–1, 235; courtship behaviour, 225 Cobitidae, biogeography, 49, 54 male competition–female cow sharks see Hexanchiformes cod choice interactions, 234–5 cowfishes, 75 Atlantic see Gadus morhua size-dependent in lake creek chub (Semotilus Pacific see Gadus macrocephalus communities, 326–7, 329, 330, atromaculatus), 259, 260, 287, coded wire tags, 191 334–5, 335 289 cods see Gadidae; Gadiformes sperm, 225, 231 critical temperature thresholds, 87 Coelacanthiformes, 19, 22, 26, 49 compressibility of water, 71 croaking gourami (Trichopsis coelacanths see Coelacanthiformes; condition factor, 108 vittata), 231 Latimeria condition indicators crucian carp (Carassius carassius), coelenterazine, 91 assessment from length–weight 286 coevolution, host–parasite, 365 relationship, 108 cruising foraging tactics, 268, 271 coho salmon see Oncorhynchus biochemical variables, 108–9 crypsis, 285 kisutch lipid–body moisture relationship, Crystallogobius linearis (crystal cohort interactions, 334 108, 109 goby), 230 competition modelling, 334–5, confusion effect, 287, 288 Ctenochaetus (girellids), 91 335 continental drift, 44 Ctenosquamata, 30 cold adaptation, 61–2 conus arteriosus, 84 cured fish, 2 aerobic metabolism, 87 convict cichlid (Cichlasoma cusk-eels see Ophidiiformes ‘antifreeze’ compounds, 81–2 nigrofasciatum), 227, 237 cutaneous gas exchange, 83–4 haemoglobin multiplicity, 87; cookie-cutter shark (Isistius cutthroat trout (Oncorhynchus Atlantic cod clinal brasiliensis), 25 clarki), 216, 286, 293, 324 polymorphism, 89 Coomassie blue staining, 206 cuttlefish (Sepia officinalis), 287 Index 395

Cuvier, Georges, 20 denticles, skin friction drag disc-shaped rajids (skates), 25 Cyclosquamata see Aulopiformes reduction, 77 dispersal, 44–6, 46 Cyclothone, 30 depensation, spawner abundance connectivity between Cymatogaster (surfperch), 151 relationship, 125 subpopulations, 209 Cypridontiformes, 19, 23, 31, 34 depensation, 2.331–2 marine fishes, 56 Cyprinidae (true minnows), 29 see also Allee effect North American postglacial, 51 biogeography, 49, 50, 51, 54 detritivores, 350 phylogeographic approaches, 46, communal breeding, 237 development, 97 47 herbivory, 91 early life history stages, 99–102 dispersalism, 46 lake species, 329; pH sensitivity, eleutheroembryonic phase, 98, 98 diurnal behavioural periodicity, 323 embryonic period, 98, 99 289–90 parasitism-related behaviour environmental sex DNA analysis change, 368 determination, 101–2 fish tissue/species identification, Cypriniformes, 19, 22, 29 rate: egg size relationship, 100, 206–7 Cyprinodon bovinus (Leon springs 100–1; temperature–incubation methods, 203–5 pupfish), 216 time relationship, 99–100, 100 uses, 202, 205 Cyprinodon variegatus (sheepshead swimbladder, 73 dogfishes see Squaliformes minnow), 216 devil ray (Mobula), 25 Dolly Varden char (Salvelinus Cyprinodontidae, biogeography, 50 diadromy, 78, 187–9 malma), 324 Cyprinodontiformes, biogeography, definitions, 176, 177, 187 dories see Zeiformes 53 exploitation during migration, drag forces Cyprinus carpio (carp), 29, 76 178 pressure, 76 biogeography, 50 facultative wanderers, 187 skin friction, 76 sex determination, 218 homing behaviour, 190, 191 swimming retardation, 76–7 see also 2.370, 2.372, 2.377, 2.382 re-establishment of populations dragonets see Gobiesociformes Cyrtocara eusinostomus (bower- following perturbation events, dragonfishes (Stomiidae), 30 building cichlid), 236, 240 193 drum snappers, 75 scale of migratory movement, 176 dynamic lift, 72 dace (Leuciscus leuciscus), 287 species distributions, 189, 193–4 Dactylopteriformes, 19, 23, 31, 35 upstream migration cababilities, East African lake species flocks, Dactyloscopidae, biogeography, 58 193 51–2 damselfish, 141, 231, 233 within-species genetic structure, Easter Island, 59 home range, 349 193 Echinorhiniformes, 19, 21, 24, 25 resource-defence polygyny, 236 see also amphidromy; anadromy; ecological niche, 321, 325 Dascyllus aruanus (damselfish), 141 catadromy body size-related ontogenetic data storage tags, 191–2 diamond leatherjacket (Rudarius shifts, 327–8 deepsea anglerfishes (ceratioids), 33, excelsus), 36 ecological study approaches, 3–4 92 diatoms, 350, 351 economically important species’ deepsea smelts see Argentiniformes fatty acid analysis, 309–10 migration, 178–9, 179 deepwater countercurrents, 180, Dicentrarchus labrax (sea bass), 278 EcoRI, 205 182, 183 climate change response ecosystem management, 301 demersal primary production, 351 prediction, 185 see also 2.90, 2.224–5, 2.293 demersal species, 346–50 migratory behaviour (seasonal Ecsenius, 56 home ranges, 349 movements), 182, 183–5, 184 ectoparasite control methods, 382 temperate, 348–50 protandrous hermaphroditism, eel see Anguilla spp. tropical, 346–8, 351 218 eels see Anguilliformes density-dependent mortality, 124–5 digestive enzymes, 90 egg-carrying behaviour, 228 at specific life history stage, digestive function, 90–1 egg guarding, 228, 229 131–3, 134, 135 piscivorous fish, 270 egg size, 151, 156 demersal juvenile stages, 140–1 Dinophyceae, fatty acid analysis, development rate relationship, estimation from long-term 309 100, 100–1 surveys, 133, 135–6 diphyllobothriasis, 377 offspring survival relationship, experimentally depleted brook Diphyllobothrium spp., 362, 377 160 trout populations, 126–8, 127 Diplostomum phoxini, 366, 368, optima, 159, 159; environment- overcompensation, 127, 128 368 specific, 161 pelagic larval stages, 136, 138–9 Diplostomum spathaceum, 362, parental care associations, 159, piscivore activities, 277 383 160 type III functional responses, 273 Dipnoi (lungfishes), 26 selection pressures, 158–9, 161 within/between-cohort data, 136 aestivation, 90 temperature effects, 161 Dentatherinidae, 34 lung/gill function, 89–90 egg trading, 239 396 Index eggs metabolic costs, 110 fast (white) muscle fibres, 76, 76 density-dependent mortality, 136, seasonal factors, 111, 112–13 fatty acid analysis 139 storage reserves, 111 food web studies, 309 development rate: size Engraulidae phytoplankton classification, relationship, 100, 100–1; anadromy, 187 309–10 temperature–incubation time catadromy, 188 featherfins see Polypteriformes relationship, 99–100, 100 Engraulis ringens (Peruvian fecundity morphology, systematic anchoveta), 2, 200 body size relationship, 97 investigations, 17 spawning aggregate parental care fitness costs, 228 ocean currents influence, 181 overexploitation, 240 recruitment relationship, 5 specific gravity reduction, 72, 73 see also 2.319, 2.320, 2.326 see also 2.296, 2.298, 2.333 yolk water content, 81 Enhydra lutris (sea otter), 313 feeding preference, 252–3 Eigenmannia virescens (knife fish), environmental sex determination, Manley–Chesson index, 252 231 101–2 feeding rate El Niño, 1 Esocidae (pikes), 30 group size effects, 255–6 see also 2.93, 2.322, 2.326 ambush behaviour, 268 ideal free distribution (IFD) elasmobranchs, 21, 24, 348 Esociformes, 19, 22, 28, 30 models, 256; predation risk age determination, 106–7 Esocoidei, biogeography, 50 incorporation, 258 dynamic lift, 72 Esox lucius (pike; northern pike), optimal foraging theory, 255, osmoregulation, 80, 81 166, 255, 269, 286, 322, 323 257 piscivory, 267 predation, 287; prey selection, 274 prey abundance relationship, 255, rectal gland, 81 see also 2.370 255 specific gravity, 72 Esox masquinongy (muskellunge), standing-stock models, 257 tissue lipids, static lift 274 temperature effects, 324, 324 contribution, 73 Esox spp. (pike), 251, 268, 270 female-defence polygyny, 236–7 see also 2.306, 2.331 estuarine habitat, 347 fighting elasmoid scales, 105, 106 Ethiopian region, freshwater threat/status signals, 231 Elassoma, 31 biogeography, 51–2 weaponry, 230 Elassomatidae, 35 Etroplinae, 52 filial cannibalism, 228–9 electric catfishes (Malapterurus), 51 Eucyclogobius newberryi, 235 fillets electric eel (Electrophorus), 29, 53 Euler–Lotka equation, 150 mislabelling/misidentification, electric knifefishes see Eurasian perch see Perca fluviatilis genetic tests, 206, 207 Gymnotiformes European bitterling (Rhodeus filming, predation assessment, electric organs, 29 sericeus), 234 278–9 electric rays, 25 European eel see Anguilla anguilla filter feeding, 268 electronic tagging, 191 European freshwater fauna, 49–50 fin clipping, 191 Electrophorus (electric eel), 29, 53 euryhaline habit, 78 fin tags, 105 Eleotridae, amphidromy, 188 Eurypterygii, 30 fineness ratio, piscivorous fish, 269 Eleotrididae, biogeography, 53, 55 Eusphrya (hammerhead shark), 24 fins elephant fishes (Mormyridae), 27, 51 Euteleostei, 28 development, 98 eleutheroembryonic phase of evolutionary taxonomy, 15, 17 swimming, 75 development, 98, 98 Exocoetidae (flying fishes), 34 fitness Elopiformes, 19, 22, 28 spawning aggregate exploitation, alternative reproductive Elopomorpha, 27–8, 73 240–1 strategies, 162–3 embiotocids, biogeography, 58 see also 2.16 body size associations, 156–7 embryonic period, 98, 99 exon-primed intron-crossing PCR habitat selection, 256, 259 cleavage phase, 98 (EPIC), 205 life history theory, 149, 150–1 eleutheroembryonic phase, 98, 98 exotic species introductions, 331, optimum egg size, 159, 159 embryonic phase, 98 331 parasite-mediated natural encounter rates hybridization/introgression, 216 selection, 366 intercohort competition parasite introductions, 376 parental care costs, 228–9 modelling, 334 piscivores, 279 flashlight fishes (Anomalopidae), 33, piscivores, 271 see also 2.86, 2.324, 2.377, 2.378, 92 endemic distribution patterns, 44–6 2.382–3 flatfishes energy budget, 109–13 eyes, 72 crypsis, 285 food consumption estimation, eyespots, 285 DNA analysis for species 116 identification, 207 food intake, 109–10 family, taxonomic, 18 see also 2.18, 2.352, 2.353 losses in faeces/nitrogenous farmed fish production see flounder, 36 excretory products, 110 aquaculture flying fishes see Exocoetidae Index 397

flying gurnards see piscivores: adaptations, 268–9; Fulton’s condition factor, 108 Dactylopteriformes schooling, 269; functional responses, 272–3 food chains, 302, 302, 303–5 search/encounter tactics, 271; temperature influence, 324 energy transfer efficiency, 303, selective feeding, 275 type I (density-independent 304 predation risk, 331 predation), 272, 273 marine fishery productivity predator choice, 253; optimal diet type II (negative density- prediction, 303–4, 304; models, 253–4 dependent predation), 273, 273, limitations of approach, 305 predator size–prey size 315 parasite transmission, 365 relationship, 334 type III (positive density- trophic levels, 303, 304 temperature effects, 323–4, 324 dependent predation), 273, 273, trophic transfers within food foraging theory, 7–8 315 webs, 310, 311, 312 see also optimal foraging theory Fundulidae, biogeography, 50 see also 2.217 Ford–Walford plot, 103, 104 Fundulus diaphanus (banded food consumption see also 2.116 killifish), 288 bioenergetic models, 115–16 forktailed catfishes see Ariidae Fundulus heteroclitus estimation methods, 115–17; Fraser–Lee back calculation (mummichog), 113 gastric evacuation rate data, method, 107 fungi, 360, 361 116–17; inputs from laboratory French grunt (Haemulon experiments, 116; stable flavolineatum), 278 Gadidae (cods), 32, 151 isotopes, 115; stomach contents freshwater biogeography, 49–55 age at maturity, 157 analysis, 115 Africa (Ethiopian region), 51–2 anadromy, 187 food deprivation response, 111, Australian region, 54–5 biogeography, 56, 344 112 Central/South America, 52–4 Gadiformes, 19, 22, 31, 32–3 seasonal variation, 112 Europe, 49–50 Gadus macrocephalus (Pacific cod), temperature influence, 110, 111 North America/Mexico, 50–1 278 food webs, 8, 302–3, 303, 305–12, Southern Asia, 54 see also 2.21 321 freshwater fish Gadus morhua (Atlantic cod), 32 diagrammatic representations, blood osmolality, 78 age at maturity, 166; growth rate 311; limitations, 310 community ecology, 321–37; relationship, 166 ecological community stability, see also 2.345, 2.346 age/size-related changes, 8 314–15 density-dependent mortality catch rate, 8 fatty acid analysis, 309 during preadult stages, 136, 138, egg size, 158, 160, 161; larval fisheries modelling, 315–16 139 survival relationship, 160 gut contents data, 305–6, 307, genetic diversity studies, 208, 208, feeding behaviour, 305 309 209 genetic analysis of stocks, 192–3, habitat choice, 261–2 osmoregulation, 78–80; gills, 78, 212–13 impact of piscivore removal by 79; kidneys, 78–80, 80 haemoglobin variation, 5, 202, fishing, 279, 301; see also 2.328, species numbers, 49, 49 213; clinal polymorphism, 2.342, 2.343, 2.344–5, 2.346, stock structure, 211 87–9 2.353–4 freshwater spiny eels larval predation, 275 interaction strength, 312–14; (Mastecembeloidei), 31, 34 lekking behaviour (spawning study approaches, 312–13 freshwater stingray (Potamotrygon), aggregations), 7, 236, 349 nutrient cycling, 332–3 80 length frequencies at low stable isotope studies frill sharks see Hexanchiformes and high abundance, 142, 143 (fingerprints), 306–8, 307, 308; frogfishes see antennariids migratory behaviour, 186, 189, advantages/disadvantages, 308, frontal zones, 350 190; scale of movement, 176 309 frozen fish, 2 offspring size/numbers, 158, 160 trophic transfers, 310, 311, 312 FST statistics, 208, 208, 209–10 population collapse, 9, 140, 200, see also 2.306 Gadus morhua (cod) stock 212; see also 2.176, 2.177, 2.180, foraging behaviour, 251–6 structure, 212, 213 2.319, 2.320, 2.332 feeding preference, 252–3, 275 Oncorhynchus keta (chum recruitment: density-dependent feeding rate: group size effects, salmon) stock structure, 211, mortality in juveniles, 132–3, 255–6; prey abundance 212 134, 135; Iceland population relationship, 255, 255 Thunnus alalunga (albacore tuna) recruitment–spawner habitat selection, 256–62 stock structure, 214 abundance relationship, 128, intercohort competition Thunnus albacares (yellowfin 128–9; population recovery modelling, 334 tuna) stock structure, 213, from low spawner abundance, light effects, 324–5 214 125; variability, 135, 138, 140 parasitism-related changes, 367, Thunnus obesus (bigeye tuna) response to fishing, 167; see also 368 stock structure, 214 2.251 398 Index

Gadus morhua (cont.) see also 2.325 see also 2.40, 2.84, 2.321, 2.321 spawning aggregate gene frequencies, statistical glochidia larvae parasites, 360, 361 overexploitation, 240; methodology, 206 glomerular filtration rate, 78 see also 2.330 genetic analysis, 200–7 Gnathostomata, 21 temperature gradient effects, 5, discrete stocks of migratory fish, Gobiesocidae, biogeography, 61 348 192–3; diadromy, 193 Gobiesociformes, 19, 23, 31, 35 vertebral numbers, 101 fish tissue identification, 206–7 Gobiidae (gobies) see also 2.21, 2.24, 2.41, 2.131, host–parasite coevolution, 365 amphidromy, 188, 193 2.131, 2.132, 2.133, 2.135, hybridization studies, 207 anadromy, 187 2.142, 2.154, 2.154, 2.155, methods, 200–5, 201, 202 biogeography, 49, 52, 54, 55, 56, 2.156, 2.160, 2.160, 2.175, population genetic studies, 58, 59, 61 2.176, 2.177, 2.184, 2.185, 207–15 catadromy, 188 2.185–6, 2.188, 2.241, 2.301, species identification, 206–7 home range, 349 2.302 statistical methods, 205–6 juvenile stage density-dependent gag grouper (Mycteroperca genetic basis of sex determination, mortality, 141 microlepis), 241 218 resource-defence polygyny, 236 Galapagos Islands, 59 genetic divergence, phylogeographic simultaneous hermaphroditism, Galaxias maculatus (inanga), 188 studies, 46, 47 239 distribution, 193 genetic diversity Gobioidei, 35 Galaxiidae bottleneck effects, 215 Gobiomorphus breviceps (upland amphidromy, 188, 193 conservation, 217 bullie), 233, 369 biogeography, 61 genetically engineered/transgenic Gobiusculus flavescens (two- catadromy, 188 fish introductions, 217 spotted goby), 8, 288 migratory larvae exploitation, 178 response to fishing, 165–6, 167, migratory larvae exploitation, 178 Galaxioidea, biogeography, 55 215–16 selective feeding, 275 Galeocerdo cuvier (tiger shark), 24 stocking programme effects, goblin shark (Mitsukurina owstoni), Galeomorphii, 24 216–17 24 game theory, 239 genetic drift, 215 goldfish (Carassius auratus), 29, 256 predator–prey model of habitat genetically engineered fish biogeography, 50 choice, 260–2, 261 introductions, 217 gonadosomatic index (GSI), 153, 154 see also 2.266–7, 2.285 genus, 18 Gondwanaland, 52, 55, 61 gape-limited predators, 274–5, Geotria australis (southern pouched Gonorynchiformes, 19, 22, 29 276–7, 286 lamprey), 187, 193 good gene models (viability gars see Lepisosteiformes Geotriidae, 21 indicator models), 231 gas exchange, 83–4 anadromy, 187 Goodeidae, biogeography, 50 air-breathing organs, 89–90 ghost pipefishes (Solenostomidae), goosefishes (lophiids), 33 environmental temperature 35 Gpi-A, 214 effects, 87–9 Gill, Theodore, 20 grades (paraphyletic taxa), 17 hypoxic environments, 89 gill-nets, 167 Grammatidae, biogeography, 58 gas fluxes in water, 82–3 see also 2.13, 2.14, 2.23–4, 2.194, grayling (Thymallus thymallus), Gasterosteidae (sticklebacks), 35 2.368, 2.370, 2.373 166 anadromy, 187 Gilliam’s rule, 259 Great Barrier Reef, 59 behavioural characteristics, 18 see also 2.234, 2.240 great white shark (Carcharodon mate choice, 233, 235 gills carcharias), 24 parasitism, 8–9 blood flow, 84 great-crested grebe (Podiceps resource-defence polygyny, 236 gas exchange, 83 cristatus), 375 Gasterosteiformes, 19, 23, 31, 34–5 modification as air-breathing green sunfish (Lepomis cyanellus), Gasterosteus aculeatus (threespine organs, 89 325, 326 stickleback), 155, 156, 231, 256, osmoregulation, 78; freshwater Greenland shark (Somniosus 286, 288, 292, 293 fish, 78, 79; marine teleosts, 81 microcephalus), 25 parasitism, 362, 363, 367 surface area, 83 grenadiers see Macrouridae Gasterosteus sp. (white ventilatory flow, 83 grey mullets see Mugiliformes stickleback), 153 Ginglymodi see Lepisosteiformes group size gastric evacuation rate, food Girardinichthys multiradiatus effect on feeding rate, 255–6 consumption estimation, (Amarillo fish), 227, 233 optimum, 256 116–17 girellids (Ctenochaetus), 91 predation risk influence, 259–60 gene flow glaciation, 51 groupers, protogynous diadromous species, 193 glass-eels, commercial expolitation, hermaphroditism, 238 freshwater versus marine fish, 209 179 grouping as prey defence, 287–9 statistical methodology, 206 global catch estimates, 304 confusion effect, 287, 288 Index 399

encounter-dilution effect, 287 Gymnocephalus cernua (ruffe), 324, hatchetfishes (Sternoptychidae), 30 group composition, 288 325 biogeography, 56 group geometry, 288–9 Gymnothorax meleagris (moray hatching, 98, 99 group size, 287–8 eel), 285 Hawaii, 59 oddity effect, 288 Gymnotiformes, 19, 22, 29 HbI, 213 growth, 5, 97–117 biogeography, 53 heart, 84 at different latitudes, 112, 113; gynogenetic species, 218 Hector’s dolphin (Cephalorhynchus growth compensation models, Gyrodactylus salaris, 376 hectori), 215 113–15, 114 helminth pathogens, 377, 378–81 early life history stages, 99–102 habitat selection, 251, 256–62 Hennig, 16, 18 energy budget, 109–13; food abiotic costs, 257–8 herbivorous fish consumption estimation, 116; food web theory, 261–2 alimentary tract, 90, 91; microbial food intake, 109–10; losses in ideal free distribution (IFD), symbionts, 91 faeces/nitrogenous excretory 256–7 energy budget, 110 products, 110; metabolic costs, parasite infection-related marine productivity, 351 110; seasonal factors, 111, changes, 369, 374 stable isotope studies, 308, 309 112–13; storage reserves, 111 predation risk, 258–60 hermaphrodites, 238–9 food deprivation response, 111, predator interference, 257 protandry, 218, 239 112 predator–prey game theory model, protogyny, 218, 238–9 length data: growth curves, 102, 260–2, 261 self-fertilizing, 239 102–3; length–weight see also 2.240, 2.241, 2.242 sequential, 238 relationships, 107–9; see also habitat structure simultaneous, 238, 239 2.189, 2.190 coral reef, 347 herring see Clupea harengus models/equations, 102–4; see also spawning areas, 349 (Atlantic herring); 2.276, 2.279–80 temperate demersal Clupeiformes mortality rate relationship, environments, 348–9 Heterandria, 53 140–1; see also 2.72 haemal arc pump, 84 Heterodontiformes, 19, 21, 24 photoperiod responsiveness, haemoglobin, 84–7, 85 Hexagrammidae, biogeography, 58 112–13 blood perfusion requirement Hexanchiformes, 19, 21, 24, 25 reproduction-related reduction, relationship, 84–5, 85 Hiodon, 50 155 multiplicity, 5, 86–7, 88, 202; Hiodontiformes, 19, 22, 27 temperature influence, 110–11, adaptive features, 87; Atlantic Hippocampus angustus, 237 111, 114 cod clinal polymorphism, Hippocampus fuscus, 237 growth curves, 102, 102–3 87–9 Hippocampus spp. (seahorses), 251 growth mortality hypothesis, 140–1 oxygen transport, 85–6, 86 monogamous breeding systems, growth rate structural aspects, 85 237 age at maturity relationship, 166 Haemulidae, biogeography, 56, 58, Hippocampus whitei, 237 alternative reproductive 344 Hippoglossus hippoglossus strategies maintenance, 164–5 Haemulon flavolineatum (French (halibut), 36, 114 life history characteristics, 156–7; grunt), 278 see also 2.65 invariants, 158; piscivores, 269, hagfishes see Myxiniformes Hippoglossus stenolepis (Pacific 270 hakes see merluciids halibut), 58 optimal foraging models, 254–5, halibut (Hippoglossus see also 2.21, 2.65, 2.70, 2.177, 255 hippoglossus), 36, 114 2.343 optimal habitat selection (m/g see also 2.65 historical biogeography, 43–62 rule), 259 halosaurs see Notacanthiformes analytical methods, 46–7, 62 parasite infection impact, 377 hamlet (Hypoplectrus nigricans), concepts, 44 response to fishing, 165, 166 239 dispersal processes, 44–6, 46 temperature effects, 324 hammerhead sharks distribution/faunal composition growth-enhanced transgenic fish (Sphyrna/Eusphrya), 24 by region, 47, 49–62; freshwater introductions, 217 handfishes (brachionichthyids), 33 regions, 49–55; marine regions, growth–maturity–longevity (GML) handling time, 254, 272, 272, 276, 55–62 plot, 103 291 vicariance processes, 44–6, 46 grunts (Haemulidae), 279 body size influence, 334 Holocentridae (squirrelfishes), 33 Gulf Stream, 180 prey morphological defences, Holocephali, 21 gulpers see Anguilliformes 286 Holostei, 26 gunnel, 75 Haplochromis spp., 331 home range, 349 Günter, Albert, 20 Haptophyceae, fatty acid analysis, homing behaviour, 189–90 guppy see Poecilia reticulata 309 hook-and-release rules, 241 gurnard (Trigla spp.), 348 Harpagiferidae, biogeography, 61 see also 2.373, 2.379–80, 2.382 400 Index

Hoplostethus atlanticus (orange Indostomidae (paradox fishes), 35 osmoregulation: freshwater roughy), 34, 166, 194, 215 Inermiidae, biogeography, 58 teleosts, 78–80, 80; marine spawning aggregate exploitation, ingestion processes, piscivores, teleosts, 81 240 268–9 structural aspects, 78, 80 see also 2.18, 2.198, 2.259, 2.265, inland catch statistics, 2, 2, 3 killifish (Rivulus marmoratus), 239 2.331 inland silverside (Menidia killifishes hornsharks see Heterodontiformes beryllina), 231 biogeography, 52 host–parasite coevolution, 365 input-matching rule, 256 molecular phylogenetic studies, hound-sharks (Triakidae), 24 interactive segregation, 325 52 human consumption of fish, 2–3, 3 interspecific interaction coefficient, kin selection, 237 human helminth pathogens, 377, 142–3 king crab (Paralithodes 378–81 introgression following species camtschatica), 207 Huxley, Thomas Henry, 3 introductions, 216 knife fish (Eigenmannia virescens), see also 2.3, 2.65, 2.66, 2.67 ion regulation, 71 231 hybridization chloride cells, 81, 82 Kudoa spp., 360 following exotic species elasmobranch rectal gland, 81 Kudoa thyrsites, 377 introductions, 216 gills: freshwater teleosts, 78, 79; Kuhliidae, catadromy, 188 following stock enhancement marine teleosts, 81, 82 Kuroshio, 180, 181 programmes, 216 kidneys, 78–9 genetic analysis, 207 specific gravity reduction, 72–3 Labridae (wrasses), 9, 251 see also 2.325 Isistius brasiliensis (cookie-cutter alternative reproductive hybridogenetic species, 218 shark), 25 strategies, 162 Hydrocyon vittatus (tiger fish), 268 island biogeography, 322 biogeography, 58, 59, 61, 344 hydrographic containment, 180 isoelectric focusing, 202 cleaning behaviour, 382; hypnids (electric rays), 25 fish tissue/species identification, exploitation in fish farms, hypo-osmotic regulation, 72 206 382–3 Hypoplectrus nigricans (hamlet), isoenzymes, systematic protogynous hermaphroditism, 239 investigations, 18 238 hypoxia, adaptive changes in oxygen isohaemoglobins, 5, 86, 87, 88, 202 resource-defence polygyny, 236 transport, 89 Atlantic cod clinal Labrisomidae, biogeography, 58 polymorphism, 87–9 Labroidei, 35 Ichthyophorus sp., 9 isolation by distance, 192 Labroides spp., cleaning behaviour, Ichthyophthirius multifilis (‘white isopod crustacean parasites, 360, 382 spot’ disease), 361–2 361 Lake Baikal, 50 Ictaluridae, biogeography, 50 Istiblennius, 56 Lake Nabugabo, 52 Icthyophonus, 377 iteroparous species, 99 Lake Tanganyika, 51, 52 ideal free distribution (IFD), 251, see also 2.28 256–7 jacks (Carangidae), 75 lake trout (Salvelinus namaycush) abiotic costs incorporation, 258 biogeography, 56, 58 functional responses, 273 input-matching rule, 256 see also 2.304 sex determination, 218 predation risks incorporation, Japanese eel (Anguilla japonica), 180 Lake Victoria, 51, 52 258–9 Jasus edwardsii (spiny lobster), 207 see also 2.6, 2.28, 2.325 standing-stock models, 257 jellynose fishes see lakes see also 2.234 Ateleopodiformes alternative highly productive illicium, 33 Jordan, David Starr, 20 states, 333–4 immobility as prey defence, 284, 285 Jordan’s rule, 101 community patterns, 322–3; immune systems, 382 juvenile stages, 98–9 extinction factors, 322, 323; inanga (Galaxias maculatus), 188 density-dependent mortality, immigration rates, 322; distribution, 193 132–3, 134, 135, 135, 136, 140–1 isolation factors, 322, 323 incubation period, 98 exploitation during migration, density-dependent mortality indices of biotic integrity (IBI), 194 179 during preadult stages, 136, 138 individual-based models, 142 growth rate, 157 habitat structure, 330–1 see also 2.228–4, 2.285 marine communities, 351 population dynamics, 334–6 Indo-Pacific see also 2.296, 2.298 size-structured community coral reef, 344 productivity, 328–30 temperate marine biogeography, Katsuwonus pelamis (skipjack species richness, 322; pH 60–1 tuna), 72, 76 sensitivity, 322–3 tropical marine biogeography, see also 2.241 temperature influence on species 59–60 kidneys distribution, 324 Indonesian archipelago, 54, 59 Antarctic teleosts, 82 Lamniformes, 19, 21, 24–5 Index 401

Lampetra spp. (lamprey), 75 leaf-fish (Monociorrhus invariant features, 157–8 osmoregulation, 78 polyacanthus), 285 offspring size/numbers trade-off, see also 2.194 leeches, 360, 361 158–62, 159 lampreys (Petromyzontidae) Leiostomus xanthurus (spot), 273 parasite infection impact, 369 anadromy, 187 lekking behaviour, 7, 236 piscivorous fishes, 269–71, 330; blood-sucking habit, 268 Gadus morhua (Atlantic cod), 7, prey diversity/alternatives, 286 environmental sex 236 predation impact, 278 determination, 218 length data reproductive costs, 154–6 gills ventilation, 83 age determination: back reproductive effort, 149, 153–4 parasitism, 360, 361 calculation, 107; resource polymorphisms, 270–1 see also Petromyzontiformes length–frequency analysis, 105; sex change/hermaphroditism, Lampridiformes, 19, 22, 31–2 measurements of hard parts, 238, 238–9 Lamprologus brichardi, 237 105, 106, 107; see also 2.72, size at maturity, 151, 152–3 lancetfishes (alepisaurids), 30 2.192, 2.279 stages, 5, 6; density-dependent lantern bass (Serranus baldwini), growth assessment: growth mortality, 131–3, 134; early 236 curves, 102, 102–3; growth/development, 99–102; lanternfishes see Myctophiformes length–weight relationships, terminology, 97–9 largemouth bass (Micropterus 107–9 variability, 151–2 salmoides), 8, 116, 258, 270, life history invariants, 158 life histories, 2.229, 2.230, 2.235–6, 322, 323, 326, 330, 332 see also 2.188, 2.204–8, 2.280–1 2.237–40 see also 2.239 length to maximum depth (fineness) see also egg size; fecundity; larval stages, 98, 99 ratio, 269 growth crypsis, 285 length–frequency analysis, 105 life span, 158 density-dependent mortality, 136, see also 2.188–208 light 138–9 length–weight relationships species distribution influence, interannual variability in condition indicator (condition 321 survival, 135, 136, 137; factor), 108 species performance effects, match/mismatch hypothesis, growth assessment, 107–9 324–5 137, 137–8, 140 Leon springs pupfish (Cyprinodon light organs, 30, 33 migration: commercial bovinus), 216 light-producing photophores, 72, 92 exploitation, 178; Leontief matrix, 314 lightfish (Vinciguerria spp.), 151 see also amphidromy Lepeophtheirus salmonis, 9 Ligula intestinalis, 9, 362, 364, 368, protopterygiolarval phase, 98 Lepidogalaxias (salamanderfish), 29 375 pterygiolarval phase, 98 Lepidogalaxias salamandroides Lindeman spine, 312 refuge use, 289–90 (salamanderfish), 55, 90 lion-fish (Pterois volitans), 75 stochastic density-independent Lepidosireniformes, 19, 22, 26 Liparidae, biogeography, 58, 62 mortality, 136, 139–40 Lepisosteidae, biogeography, 50 lipase, 90 see also 2.238–9, 2.296, 2.298, Lepisosteiformes, 19, 22, 27 lipids 2.300 Lepomis cyanellus (green sunfish), body moisture relationship, 108, Lates calcarifer (barramundi) 325, 326 109 catadromy, 188 Lepomis gibbosus (pumpkinseed energy reserves, 111 protandrous hermaphroditism, sunfish), 153, 157, 325, 327 reproductive costs, 156 218 Lepomis macrochirus (bluegill stable isotope studies, 307 Lates nilotica (Nile perch), 279, 331 sunfish), 8, 153, 255, 255, 257, static lift provision, 73 see also 2.6, 2.193, 2.325, 2.368 258, 325–7, 330, 332 littoral fish faunas Latimeria, 80, 81 Lepomis spp. (sunfishes), 325 temperate, 344–5 Latimeria chalumnae, 26 Leporinus, 218 tropical, 342, 343–4 see also 2.331 leptocephalus larva, 188 live bearers (Xiphophorus), 53 Latimeria menadoensis, 26 Leuciscus leuciscus (dace), 287 lizard-fishes see Aulopiformes latitudinal variation life histories, 149–68 loaches see Cypriniformes biodiversity trends, 342–3, 343 age at maturity, 151, 152–3 locomotion, aquatic adaptations, 71, growth, 112, 113; compensation alternative reproductive 75–7 models, 113–15, 114; seasonal strategies, 162–5 Lolio vulgaris (squid), 287 responses, 112–13 critical period hypothesis, 131 longfin eel (Anguilla dieffenbachii), parasite communities, 363–4 definition, 149–50 188 temperate demersal temperature effects of fishing, 165–7, 167; see Lophiiformes (anglerfishes), 19, 23, gradients, 348 also 2.331, 2.335–6 31, 32, 92 vertebral numbers, 101 fitness relationship, 149, 150–1 luring behaviour, 268 Laurasia, 50 growth rate, 156–7 Lophius spp., 251 Ldh-3, 213 habitat use changes, 178 biogeography, 56, 57 402 Index

Loricaridae (armoured catfish), 257, marine protection areas (MPAs), 7, Menidia beryllina (inland 332 279 silverside), 231 luciferases, 91 food webs, 315 Menidia menidia (Atlantic luciferins, 91 see also 2.224, 2.293–311 silverside), 113, 155 lungfish (Protopterus), 51 mark–recapture data, age meristic characters, temperature aestivation, 90 determination, 104–5 influences, 101 lungfishes see Dipnoi Mastecembeloidei (freshwater spiny Merlangius merlangus (whiting), luring behaviour, 268 eels), 31, 34 202 Lutjanidae (snappers) match/mismatch hypothesis, see also 2.156, 2.343, 2.369 biogeography, 344 137–8, 140, 141 Merluccius productus (Pacific hake), catadromy, 188 mate choice, 225 156 Lutjanus campechanus (red copying, 234 merluciids, biogeography, 56 snapper), 206, 207 direct selection processes, 232, meta-analysis, 144 Luxilus, 207 233 compensatory reserve, 129–30 environmental influences, 227 multispecies interactions, 143 mackerel see Scomber scombrus good gene models (viability recruitment–environment mackerel sharks see Lamniformes indicator models), 231 correlations, 140 Macquaria, 207 male competition–female choice metamorphosis, 98, 99, 351 macroalgae, 350 interactions, 234–5 Mexico, 50–1 Macrouridae, 32, 49 mate sampling, 233 microbial symbionts, herbivorous biogeography, 61 ornamental traits, 231 fish, 91 Madagascar, 52 search costs, 233 Micromesistius poutassou Makaira indica (black marlin), 35 sexual selection, 230, 231–5 (blue whiting), 106 Malapterurus (electric catfishes), spawning sites, 234 see also 2.19 51 maternal care, 228 Micromyzon akamai, 53 Maldives, 60 evolutionary aspects, 230 Micronesia, 59 Mallotus villosus (capelin), 160 fecundity cost, 229 Micropterus dolomieui see also 2.16, 2.19, 2.184, 2.236, see also parental care (smallmouth bass), 154, 241 2.241, 2.242–3 mating patterns, 236–8 Micropterus salmoides (largemouth management strategies, 9–10 resource distribution bass), 8, 116, 258, 270, 322, 323, mangrove habitat, 347 relationship, 236 326, 330, 332 Manley–Chesson index, 252 see also alternative reproductive see also 2.239 Manta (manta ray), 25 strategies; breeding systems microsatellite DNA analysis, 204 Mardaciidae, 21 Mauritius, 60 genetic diversity studies, 208, 209, marginal predation, 289 maximum reproductive rate 210 marine biogeography, 55–62 constancy of annual reproductive stock structure genetic analysis: Antarctic, 61–2 rate, 130 bigeye tuna (Thunnus obesus), antitropical/antiequatorial definition of exploitation limits, 215; chum salmon distributions, 60 126 (Oncorhynchus keta), 211, 212; eastern North Pacific, 58 estimation methods, 124; cod (Gadus morhua), 192–3, Indo-Pacific: temperate, 60–1; spawner–recruit plots, 126 212; yellowfin tuna tropical, 59–60 experimentally depleted brook (Thunnus albacares), 214 Mediterranean, 56–7, 57 trout (Salvelinus fontinalis) Midas cichlid (Cichlasoma North Atlantic, 56–7, 57 populations, 127, 129 citrinellum), 237 tropical eastern Pacific, 58–9 Iceland cod populations, 128–9 migration, 6–7, 99, 175–95 tropical western Atlantic, 57–8 spawner–recruit relationship, cost–benefit trade-offs, 189, 190; marine catch statisitics, 1–2, 2, 3 124, 129 metabolic costs, 186; parasite marine community ecology, 341–53 striped bass (Morone saxatilis), infection risk, 375 demersal species, 346–50 129 definitions, 175–6 pelagic species, 349–50 medaka (Oryzias latipes), 227 diadromous species see recruitment dynamics, 351–2 Mediterranean, 344 diadromy marine fish marine biogeography, 56–7, 57 discrete stock genetic analysis, egg adaptations, 81 Megalops atlanticus (tarpon), 28 192–3 genetic diversity studies, 208, 208, Megashasma pelagios (megamouth economically important species, 209 shark), 25 178–9, 179 osmoregulation, 80–1 Meiacanthus atrodorsalis, 285 exploitation of spatiotemporal species numbers, 49, 49 Melanotaenia trifasciata, 54 concentrations, 178–9, 194 stock structure, 211 Melanotaeniidae, 34 facultative wanderers, 187 marine food webs see food webs biogeography, 55 genetic connectivity between Index 403

subpopulations, 208, 209; FST mollusc larvae parasites, 360, 361 Myxiniformes, 19, 21, 21 methods, 208, 208, 209–10 Monacanthidae, biogeography, 61 circulatory system, 84 highways, 349 monkfishes (lophiids), 33 egg size, 151, 158 impact on abundance surveys, Monociorrhus polyacanthus osmoregulation, 80–1 194 (leaf-fish), 285 Myxobolus cerebralis, 362 navigation/homing, 189–90; monogamy, 237 learning through ‘dummy run’ environmental influences, 227 15N isotope studies migrations, 185, 190; oriented mooneyes see Hiodontiformes food web anaylsis, 306 swimming, 189 moray eel (Gymnothorax see also 2.219 nutrient translocation, 333; meleagris), 285 nanoplankton, 304 see also 2.349 Mordaciidae, anadromy, 187 narcinids (electric rays), 25 oceanodromous species see Mormyridae (elephant fishes), 27, 51 narkids (electric rays), 25 oceanodromous fishes Morone saxatilis (striped bass), 76, navigation, 189–90 rationale, 178 113, 160, 274, 366 Nearctic region freshwater fauna, scale of movements, 157, 176 recruitment–spawner abundance 50–1 sex-specific fisheries, 241 relationship, 129, 129 needlefishes (Belonidae), 34 species distribution, 192–4 mortality stalking behaviour, 268 study techniques, 191–2; stable density-dependent see density- negative frequency-dependent isotope markers, 306 dependent mortality selection, 163, 163 terminology, 176–8 life history invariant, 158 Nelson, G.J., 20 see also 2.38, 2.205–6, 2.206, see also 2.72, 2.189, 2.190 nematodes, 360, 361, 362–3, 368 2.207, 2.235, 2.235–6, 2.240–1, mouth, 90 human pathogens, 377, 381 2.265, 2.267, 2.281, 2.282, piscivorous fish, 268, 270; prey Neoceratodus forsteri 2.298, 2.300, 2.330 size limitation, 274–5, 276–7, (Queensland lungfish), 55 mimicry, 285–6 286 Neochanna, 90 evolutionary aspects, 288 mouth brooding, 158, 228, 230 Neocirrhites armatus, 237 minnow (Phoxinus phoxinus), 231, MseI, 205 Neolamprologus pulcher, 237 256, 286, 366 mucous, skin friction drag Neoselachii, 24 minnows (Phoxinus spp.), 251 reduction, 77 Neoteleostei, 28, 30 asexual species, 237–8 mudminnows (Umbridae), 30 Nerophis ophidion (pipefish), 228, refuge use, 290 Mugil spp. (mullets), 287 235 mitochondrial DNA analysis, 18, Mugilidae, 31 New Guinea, 54–5, 59 209 catadromy, 188 New Zealand, 61 fish tissue/species identification, Mugiliformes, 19, 23, 31, 34 Nile perch (Lates nilotica), 279, 331 207 Müllerian mimicry, 285 see also 2.6, 2.193, 2.325, 2.368 hybridization studies, 207 multispecies models, 8 nocturnal activity, 290, 291 loss of genetic diversity due to see also 2.213–14, 2.221–2, 2.222 feeding, 324 fishing, 215–16 multispecies virtual population non-reproducing helpers, 237 methods, 203 analysis (MSVPA), 8 North America stock structure genetic analysis: functional responses, 273 Atlantic region, 344 albacore tuna (Thunnus see also 2.214, 2.285 freshwater fauna, 50–1 alalunga), 214; bigeye tuna mummichog (Fundulus Pacific region, 344 (Thunnus obesus), 214; chum heteroclitus), 113 North Atlantic Drift, 180, 183 salmon (Oncorhynchus keta), Muraenidae, biogeography, 56, 58, North Atlantic 211, 212; cod (Gadus morhua), 59 (Mediterranean–Atlantic), 212, 213; yellowfin tuna muscle fibres, 76, 76 56–7, 57, 344 (Thunnus albacares), 214 metabolism during swimming, 77 North Equatorial current, 181 mitochondrial DNA inheritance, muscle tissue, species identification North Pacific, 58 203, 203 methods, 206 North Sea plaice see Pleuronectes Mitsukurina owstoni (goblin shark), musculature, systematic platessa 24 investigations, 17 North Sea tidal currents, 182 Mobula (devil ray), 25 muskellunge (Esox masquinongy), northern pike see Esox lucius Mochokidae, biogeography, 51 274 northern squawfish (Ptychocheilus Mogurnda, 52 mycobacteria, 207 oregonensis), 273 Mola mola (ocean sunfish), 36, 158 Mycteroperca microlepis (gag Notacanthiformes, 19, 22, 28 molecular clocks, phylogeographic grouper), 241 Notobranchius sp., 151 studies, 46, 47 Myctophiformes, 19, 22, 30, 31 Nototheniidae, biogeography, 61 molecular techniques, 7 myotomes, 75–6 Notothenioidei, 35 Molidae, 151 slow/fast muscle fibres, 76, 76 Antarctic, 61, 87 404 Index numerical taxonomy (phenetics), density-dependent mortality ostraciform swimming, 75 15–16 during preadult stages, 136 Ostraciidae, biogeography, 61 nurse shark, 75 distribution, 193 otolith sampling, 105, 106 nutrient cycling, 332–3 genetic diversity conservation, recruitment studies, 133 see also 2.349 217 see also 2.72, 2.280, 2.347 maximum reproductive rate Otophysi, 29 oarfish (Regalecus glesne), 32 calculation, 126 outbreeding depression, 216 oarfishes see Lampridiformes migration: anadromy, 187; overcompensation, 127, 128 ocean currents, 180, 180, 181 homing behaviour, 190; scale of oviparous species, deepwater countercurrents, 180, movement, 176 growth–maturity–longevity 182 recruitment variability, 133 (GML) plots, 103 influence on migratory behaviour, Y-chromosome-specific probes, ovoviviparity, 98 180–1 218 Oxycirrites typus, 237 ocean productivity, 303–4, 304 see also 2.349 oxygen availability, 82–3 ocean sunfish (Mola mola), 36, 158 Oncorhynchus tshawytscha temperature influence, 89 oceanodromous fishes, 179–87 (chinook salmon), 167, 187, 193 oxygen transport, 85–6, 86 definition, 176 see also 2.379 Bohr effect, 85, 87 water current contributions to operational sex ratio, 226, 235 environmental temperature fish movements, 179–82, 189, environmental influences, 227 effects, 87–9 190 Ophichthyidae, biogeography, 56 hypoxic environments, 89 Odacidae, 91, 351 Ophidiiformes, 19, 22, 31, 32 Root effect, 85 oddity effect, 288 Ophioblennius atlanticus (redlip Odontaspis taurus (sand tiger), 151 blenny), 233 Pachypanchax, 52 oesophagus, 90 optimal foraging theory, 251, 253–4, Pacific cod (Gadus macrocephalus), offspring size/numbers trade-off, 326 278 158–62, 159 growth rate prediction, 254–5, 255 see also 2.21 Oncorhynchus apache prey size, 276 Pacific hake (Merluccius productus), (Apache trout), 216 see also foraging theory 156 Oncorhynchus clarki (cutthroat optimal habitat selection, 251, 259, Pacific halibut (Hippoglossus trout), 216, 286, 293, 324 326 stenolepis), 58 Oncorhynchus gorbuscha orange roughy see Hoplostethus see also 2.21, 2.65, 2.70, 2.177, (pink salmon), 187, 216 atlanticus 2.343 Oncorhynchus keta (chum salmon), order, taxonomic, 18, 19 Pacific rockfish (Sebastes spp.), 207 187 Orectolobiformes, 19, 21, 24 see also 2.19, 2.31, 2.198, 2.300, stock structure genetic analysis, ornamental traits 2.327 211–12, 212 condition-dependent expression, Pacific salmon see Oncorhynchus Oncorhynchus kisutch (coho 234 spp. salmon), 167 females, 234 paddlefish (Polyodon spathula), 50, habitat selection, 258 good gene models (viability 231 mating competition, 235; indicator models), 231–2 paddlefish (Polyodon spp.), 251 hatchery populations, 240, 240 sexual selection, 230, 231, 233–4, paddlefishes see Acipenseriformes transgenic growth hormone- 366 pair bonds, 225 enhanced, 217 ornithine decarboxylase as environmental influences, 227 Oncorhynchus mykiss (rainbow condition indicator, 109 panbiogeography, 46 trout), 29 Oryzias latipes (medaka), 227 pancreatic tissue, 90 hybridization, 216 Osmeridae (northern smelts), 193 Pantodon (butterflyfish), 51 swimming adaptations, 76, 77 anadromy, 187 Panulirus marginatus (spiny see also 2.372, 2.377, 2.381, 2.382, osmoregulation, 71–2, 78–81 lobster), 215, 313 2.383 freshwater, 78; gills, 78, 79; Paracanthopterygii, 31, 32 Oncorhynchus nerka kidney function, 78–80, 80 paradox fishes (Indostomidae), 35 (sockeye salmon), 273 sea water, 80–1 Paralichthys lethostigma (southern egg size, 160 Ostariophysi, 28–9 flounder), 273 life history characteristics, 125 alarm substances/alarm reaction, Paralithodes camtschatica nutrient translocation, 333 17–18, 28 (king crab), 207 pelagic larval stage mortality, 139 Osteichthyes, 25–6, 71 parasites, 8–9, 359–84 recruitment–spawner abundance Osteoglossiformes, 19, 22, 27 behavioural avoidance relationship, 125–6, 126 Osteoglossomorpha, 27 mechanisms, 382 see also 2.77, 2.77–8, 2.241 biogeography, 50 biological tagging, 383 Oncorhynchus spp. (Pacific osteological characters, systematic communities within hosts, 363–5; salmon), 99, 124, 332 investigations, 17 marine vs. freshwater species, Index 405

364–5; tropical vs. temperate protogynous hermaphroditism, perches see Perciformes fishes, 363–4 238 Percichthyidae control, 382–3; biological in fish see also 2.304 biogeography, 55 farms, 382–3; chemical, 383; parsimony analysis catadromy, 188 ectoparasites, 382; area cladogram construction, 47 Percidae intermediate host removal, 383; endemicity, 46 freshwater biogeography, 49, 50 vaccines, 383 passive integrated transponders (PIT lakes, 328, 329, 329 economic importance, 374, tags), 105 Perciformes, 19, 23, 31, 35–6, 91 376–7 paternal care, 228, 235 Percoidei, 35–6 host behaviour change, 367–9, evolutionary aspects, 229, 230 Percopsiformes, 19, 22, 31, 32 370–3, 374; antipredator fecundity cost, 229 biogeography, 50 defences, 367, 369; habitat see also parental care Perichthyidae, anadromy, 187 choice, 369, 374; mechanisms, pathogen identification, molecular Peruvian anchovetta see Engraulis 368–9; reproductive behaviour, methods, 207 ringens 374; swimming performance, Patterson, 20 Petersen discs, 191 367, 368, 369 peacock blenny (Salaria pavo), 235 Petromyzontidae, 21 host immune defences, 382 peacock wrasse (Symphodus tinca), anadromy, 187 host population ecology, 374–6; 164 biogeography, 49 anthropogenic introductions, pearlfishes see Ophidiiformes Petromyzontiformes, 19, 21, 21, 71 376; distribution patterns, Pegasidae (sea moths), 35 Phallostethidae, 34 374–5; impact of environmental pelagic eggs pharynx, 90 change, 375, 376; population density-independent mortality, phenetic classifications, 15, 16 structure, 375 136, 139 grades (paraphyletic taxa), 17 host–parasite coevolution, 365, ocean currents influence, 181 phenograms, 46 367 specific gravity reduction, 72, 73 phenotype matching, 288 human health implications, 377, yolk water content, 81 Philippines, 59 378–81 pelagic larvae photic zone adaptations, 72 information sources, 361 coral reef fishes, 346, 347, 351 Photobacterium fischeri life-cycle variation, 360; direct ocean currents influence, 181 (Vibrio fischeri), 92 transmission, 360–2; indirect specific gravity reduction, 72, 73 Photobacterium leio gnathi, 92 transmission, 362–3 pelagic species Photobacterium phosphoreum, 92 natural selection agents, 366 habitat, 349–50 photoperiod, growth response, pollution detection, 383–4 shark tissue lipids, static lift 112–13 reproduction-related infection contribution, 73 Phoxinus eos (redbelly dace), 290 risk, 156 see also oceanodromous fishes Phoxinus eos-neogaeus complex, sexual selection agents, 231–2, Pempheridae, 348 218 233, 366 pepsin, 90 Phoxinus phoxinus (minnow), 231, taxonomy, 360, 361 Perca, 36 257, 286, 366 tolerance, 382 Perca flavescens (yellow perch), 323 Phoxinus spp. (minnows), 251 parasitic catfishes (candirús), 53 density-dependent mortality asexual species, 237–8 Paratilapia, 52 during preadult stages, 136, 138 refuge use, 290 parental care, 157, 225, 227, 228–30, size-specific fecundity, 154 phylogenetic relationships, 4, 235, 237, 241 sperm numbers/testes size, 231 15–36 alloparents, 237 Perca fluviatilis (Eurasian perch), 8, apomorphic/plesiomorphic communal breeding, 237 18 characters, 16 definition, 228 cannibalism, 336, 336 fish diversity, 20–1; extant orders, direct benefits of mate choice, density-dependent mortality 19, 21–3, 21–36 232, 233 during preadult stages, 136 study methods, 15–17, 16 diversity, 228 group foraging behaviour, 269 phylogeography, 46–7 egg size associations, 159, 160 lake size-structured phylum, 18 energy costs, 156 communities, 329–30 physiologically structured models, filial cannibalism, 228–9 pH sensitivity, 323 322, 334–6, 337 fitness costs, 228–9 predation, 287 phytoplankton, 350 non-reproducing helpers, 237 selective feeding, 275 fatty acid analysis, 309–10 predation pressure influence, 227 size-dependent interactions, 326; primary production cycles: refuge use in prey, 292 roach (Rutilus rutilus), 326, stratified zones, 350–1; uniparental/biparental, 229; 328, 329–30, 331 upwelling areas, 350 evolutionary aspects, 229, 230 sperm numbers/testes size, 231 trophic cascade model, 277–8 parrotfish (Sparisoma spp.), 308 temperature-related feeding rate, pike (Esox spp.), 251, 268, 270 parrotfishes (Scaridae), 91, 251 324, 324 pike (northern pike) see Esox lucius 406 Index pikes (Esocidae), 30 Plecoglossus altivelis (ayu), 178 Polypteridae, biogeography, 51 ambush behaviour, 268 amphidromy, 188 Polypteriformes, 19, 22, 26 Pinguipedidae, amphidromy, 188 Plectropomus areolatus Pomacentridae, biogeography, 58–9 pink salmon (Oncorhynchus (coral trout), 241 Pomacentrum moluccensis, 351 gorbuscha), 187, 216 plesiomorphic characters, 16 Pomatomus saltatrix (bluefish), 76, pipefishes, 35, 251 Plesiopidae, biogeography, 56, 61 268–9 monogamous breeding systems, Pleuronectes ferruginea prey selection, 274 237 (yellow flounder), 157 Pomatoschistus microps see also Syngnathidae Pleuronectes platessa (plaice), 156 (common goby), 156, 228 piranha (Serrasalmus spp.), 268, 269 growth rate–age at maturity Pomatoschistus minutus pirarucú (Arapaima gigas; relationship, 166 (sand goby), 155, 227, 233 arapaima), 27, 53 migratory behaviour, 180, 182, pop-up tags, 183, 192 pirate perches (Aphredoderidae), 32 185–7, 186, 189, 190, 241; population dynamics see also Percopsiformes scale of movement, 176 experimental stocking studies, piscivorous birds, parasite response to fishing, 167 139 infection/transmission, 362, tagging methods, 191 lake communities, 334–6 375 see also 2.18, 2.353 physiologically structured piscivorous fishes, 267–80 Pleuronectidae, 151 models, 334–6, 337 cannibalism, 278 biogeography, 58, 61, 344 see also 2.74–8, 2.77, 2.77–8, definitions, 267 catadromy, 188 2.277, 2.285 feeding behaviour, 268–9 Pleuronectiformes, 19, 23, 31, 36 population genetics, 207–15 field study methods, 278–9 Plotosidae, biogeography, 55 anadromous species, 208, 208 impact of fishing, 279, 301 Podiceps cristatus (great-crested genetic diversity, 207–10; ingestion processes, 268–9 grebe), 375 connectivity between introductions, 279 Poecilia formosa (sailfin molly), 218 subpopulations, 208, 209; lakes community productivity, Poecilia latipinna, 218 conservation, 217; effects of 329, 329 Poecilia reticulata (guppy), 153, 227, fishing, 215–16; effects of life histories, 269–71, 330; 233, 256, 258, 278 stocking, 216–17; FST methods, resource polymorphisms, 270–1 grouping as prey defence, 287 208, 208, 209–10 morphological adaptations, 268; mate choice experiments, 232, stock structure analysis, 210–15; length to maximum depth 232 albacore tuna (Thunnus (fineness) ratio, 269; prey size parasitism resistance, 382 alalunga), 214; bigeye tuna limitation, 274–5 predator inspection behaviour, (Thunnus obesus), 214, 215; ontogenetic niche shifts, 328, 329; 284 chum salmon (Oncorhynchus onset of piscivory, 269–70 Poeciliidae, biogeography, 50, 53 keta), 211–12, 212; cod (Gadus population regulation effects, Poeciliopsis, 218 morhua), 212–13; yellowfin 277–8 Poeciliopsis lucida, 218 tuna (Thunnus albacares), predation components, 271–4; Poeciliopsis monacha, 218 213–15, 214 functional responses, 272–3, Poeciliopsis monacha-lucida population oscillations, 335 273; numerical responses, 274 biotype, 218 see also 2.77–8 predator size–prey size Poeciliopsis spp. (topminnow), porcupine fish, 286 relationships, 275–7, 276; asexual species, 237–8 Porichthyinae, 33 maximum prey size, 276–7 pollution, 5 Porichthys notatus (plainfin prey type/size selection, 274; sewage effluent, 376 midshipman), 163 behavioural mechanisms, thermal, 376 Port Jackson sharks see 274–5 use of parasites in monitoring, Heterodontiformes primary/secondary piscivory, 267 383–4 potamodromous fishes, 176 profitability estimation, 272 polygyny Potamotrygon (freshwater stingray), schooling behaviour, 269 environmental influences, 227 80 trophic dynamic effects, 331–2 female-defence, 236–7 potential reproductive rate, 226, 227 see also 2.156, 2.158–9, 2.343, resource-defence, 236 predation 2.378 polymerase chain reaction, 203–4 age-related changes, 8 placoid scales, 106–7 fish tissue/species identification, ambush tactics, 268, 271 plaice see Pleuronectes platessa 207 components, 271–4 plainfin midshipman see also exon-primed intron- encounter rate estimation, 271 (Porichthys notatus), 163 crossing PCR (EPIC) field study methods, 278–9 planktivores, 328 Polymixiiformes, 19, 22, 31, 32 foraging behaviour see foraging plastic flag tags, 105 Polyodon spathula, 50, 231 behaviour Platycephalidae, biogeography, 61 Polyodon spp. (paddlefish), 251 functional responses, 272–3, 273 platyhelminths, 360, 361 Polyodontidae, biogeography, 51 habitat structure effects, 330–1 Index 407

handling time, 272 body size–predator size Prototroctidae, amphidromy, 188 lake species richness impact, 323 relationships, 275–7, 276, protozoan parasites, 360, 361, 361, learning, 254, 273 335–6; maximum prey size, 377 marginal, 289 276–7 Pseudochromidae, biogeography, numerical responses, 274 capture, 254, 271–2, 275 56, 59 prey capture, 254, 271–2, 275 crypsis, 285 Pseudomugilidae, 34 prey detection, 254, 271, 275 defences, 275, 284–93, 285; Pseudoterranova sp., 377 prey preference, 252–3 grouping, 287–9; immobility, pseudotrichonotids, 30 prey switching behaviour, 273, 284, 285; impact of parasite Pterois volitans (lion-fish), 75 274, 315 infection, 367, 369; Pteromyzon marinus (sea lamprey), profitability estimation, 272 morphological, 275, 286; 158, 176 random (opportunistic) feeding, phenotype matching, 288 pterygiolarval phase, 98 274 kin associations, 286 Ptychocheilus oregonensis search tactics, 271; frequency- piscivore selective (specialist) (northern squawfish), 273 dependent, 288 feeding, 274–5 Ptychochromis, 52 selective (specialist) feeding, 274 piscivore switching behaviour, pufferfish, 286 size-dependent, 8, 327; body 273, 274, 315 pumpkinseed sunfish (Lepomis size–prey size relationships, predator inspection/recognition, gibbosus), 153, 157, 325, 327 275–7, 276, 335–6; maximum 284 Pycnopodia helianthoides, 313 prey size, 276–7 refuge use, 273, 289–93; trade-offs, pygmy swordtail (Xiphophorus stages, 253 290, 290–3 nigrensis), 163 see also piscivorous fishes; see also predation risk/predation pyloric caeca, 90 predator–prey interactions pressure; predator–prey predation risk/predation pressure, interactions Queensland lungfish 258–60 prey tethering experiments, 278 (Neoceratodus forsteri), 55 density-dependent mortality, 277; pricklefishes see see also 2.156, 2.158–9, 2.159, Stephanoberyciformes rainbow trout (Oncorhynchus 2.184 primary freshwater fishes, 44 mykiss), 29 diurnal behavioural periodicity, primary production hybridization, 216 289–90 benthic, 351 swimming adaptations, 76, 77 group size-related changes, estimation from food chain, see also 2.372, 2.377, 2.381, 2.382, 259–60 303–4, 304 2.383 grouping as prey defence, 287 stratified oceanic zones, 350–1 rainbow trout (Salmo gairdneri) growth mortality hypothesis, upwelling areas, 350 predation risk assessment, 278 140–1 Prionace glauca (blue shark), 77, size-dependent habitat structure, 331 189 predation/competition, 327 mate choice influence, 233 Prisoner’s dilemma, 239 rainbow-fishes see Atheriniformes parental care fitness costs, 228 pristids (sawfishes), 25 Rajidae, biogeography, 61 risk reduction strategies, 284, 285 Pristiophoriformes, 19, 22, 24, 25 Rajiformes, 19, 22, 24, 25 sexual selection influence, 227 productivity ram suspension (filter) feeding, 268 spatial refugia, 289 estimation from marine food randomly amplified polymorphic temperature effects, 293 chain, 303–4, 304, 305; DNA (RAPD), 204–5 predator interference, 257, 286 limitations of approach, 305 fish tissue/species identification, predator–prey interactions, 7 species distribution influence, 207 cyclic abundance changes, 278 321, 328–9 rate of increase, 150 food web relationships, 313, see also primary production Ratsirakea, 52 313–14, 314; top predator promiscuous group-spawning, 236 ray-finned fishes see Actinopterygii removal by fishing, 279, 301; Protacanthopterygii, 28, 29 rays see Rajiformes trophic dynamic effects, 331–2 protandrous hermaphroditism, 218, reconciled phylogenetic trees, 47 functional responses, 272–3, 273 239 recruitment, 5, 6, 123–44 game theory model of habitat protein definitions, 123 choice, 260–2, 261 reproductive costs, 156 juvenile demersal stages, 140–1 numerical responses, 274 somatic growth, 111 marine communities, 351–2 pelagic larval variability in protein electrophoresis, 200, 202 modelling approaches, 124–5, survival, 137, 137 fish tissue identification, 206 141–2; individual-based, 142; population dynamics, 277–8 protogynous hermaphroditism, 218, see also 2.116, 2.117, 2.118, see also 2.79–80, 2.213, 2.285, 238–9 2.119, 2.243, 2.276, 2.283 2.342–3, 2.345 Protopterus (lungfish), 51 multispecies interactions, 142–3 prey aestivation, 90 pelagic stages, 136–40; prediction alarm substances, 286 protopterygiolarval phase, 98 from environmental data, 140 408 Index recruitment (cont.) resource polymorphisms, 270–1 population oscillations, 335 prediction problems, 140, 141; resource-defence polygyny, 236 temperature-related feeding rate, see also 2.172, 2.173 respiratory function 324, 324 replacement rate estimation at adaptations, 82–90; swimming, 77 low abundance, 126 parasitism-related changes, 368 sailfin molly (Poecilia formosa), 218 spawner abundance relationship, see also gas exchange; oxygen St Peter’s fish (Sarotherodon 124–9; brook trout (Salvelinus transport galilaeus), 228 fontinalis), 126–8, 127; Iceland restriction enzymes, 203, 205 salamanderfish (Lepidogalaxias), cod (Gadus morhua), 128, restriction fragment length 29, 55, 90 128–9; meta-analysis, 129–30; polymorphism (RFLP), 204 Salangidae, anadromy, 187 sockeye salmon cod (Gadus morhua) stock Salaria pavo (peacock blenny), 235 (Oncorhynchus nerka), 125–6, structure analysis, 212 salinity 126; stock-recruitment models, fish tissue/species identification, frontal zone boundaries, 350 124–5; striped bass (Morone 207 phenotypic sex influence, 101 saxatilis), 129, 129 mitochondrial DNA analysis, 203 salt tolerance of fish taxa, 44 variability, 6, 131, 132, 132, 133; pathogen identification, 207 Salmo gairdneri (rainbow trout) environmental factors, 139–40; Retropinnidae, anadromy, 187 predation risk assessment, 278 life history stage relationships, Reynolds number, 76, 77 size-dependent 132–3, 134, 135, 135, 139, 154; Rheocles, 52 predation/competition, 327 see also 2.78, 2.79, 2.183, 2.238, Rhincodon typus (whale shark), 24, Salmo salar (Atlantic salmon) 2.309 151, 251 alternative reproductive rectal gland, 81 Rhinicthys atratulus (blacknose strategies, 162, 163, 164; Red Sea, 60 dace), 258 growth-rate thresholds for male red (slow) muscle fibres, 76, 76 Rhodeus, 218 parr maturation, 164, 164–5; red snapper (Lutjanus Rhodeus sericeus (European management implications, 165 campechanus), 206, 207 bitterling), 234 anadromy, 187, 193 redbelly dace (Phoxinus eos), 290 rhynids (shark rays), 25 density-dependent mortality redlip blenny (Ophioblennius ricefishes (Adrianichthyidae), 34 during preadult stages, 136 atlanticus), 233 Richardsonius balteatus distribution, 193 redside shiner (Richardsonius (redside shiner), 327 ectoparasite control: cleaner fish, balteatus), 327 Ricker stock-recruitment model, 382; resistant strain refuge use, 289–93, 331 124, 125, 127, 128, 129 introduction, 217 habitat choice see habitat see also 2.118, 2.121, 2.178, 2.179, egg size, 160 selection 2.228, 2.272, 2.302 homing behaviour, 190 spatial refugia, 289, 290–1 river bullhead (Cottus gobio), 228, life history characteristics, 151, stress reduction, 293 234 157; age at maturity variability, temporal refugia, 289–90, 291 Rivulus marmoratus (killifish), 239 153 trade-offs, 290, 290–3; benefits, RNA:DNA ratio as condition refuge use, 289, 290, 291, 293 292–3; costs, 290–2 indicator, 108–9 response to fishing, 167 see also 2.222–3 RNA:protein ratio as condition soft flesh disease, 377 refugia, North American freshwater indicator, 109 see also 2.40, 2.324, 2.330 fauna, 51 roach see Rutilus rutilus Salmo trutta (brown trout), 29, 87 Regalecus glesne (oarfish), 32 rock bass (Ambloplites rupestris), Arctic char competitive Regan, C. Tate, 20 287, 289 interactions, 325 reproductive behaviour, 7, 9, 99, Root effect, 85 biogeography, 50 225–42 rotenone, 376 chasing behaviour, 268 fisheries exploitation, 239–41 round-nosed grenadier density-dependent mortality life histories, 238–9 (Coryphaenoides rupestris), 194 during preadult stages, 132, 134, parasite infection-related Rudarius excelsus (diamond 135, 136 changes, 374 leatherjacket), 36 introgression following stock reproductive costs, 154–6 ruffe (Gymnocephalus cernua), 324, enhancement programmes, 216 experimental approaches, 155–6 325 life history characteristics, 157; quantification, 154–5 Rutilus rutilus (roach), 8 egg size, 161; invariants, 158 reproductive effort, 149, 153–4 antipredator behaviours, 275 light-related feeding behaviour, definition, 153 lake size-structured 324 egg size evolution, 159 communities, 329, 330 spatial refugia, 289 response to fishing, 165, 166 Ligula intesinalis infection, 375 temperature-related performance, requiem shark (Carcharhinus), 24 perch size-dependent 324 requiem sharks see interactions, 326, 328, 329, 330, see also 2.372, 2.378, 2.381 Carcharhiniformes 331 Salmo trutta fario, 216 Index 409

Salmo trutta marmoratus, 216 Scaridae (parrotfishes), 91, 251 Sebastes spp. (Pacific rockfish), 207 salmon protogynous hermaphroditism, see also 2.19, 2.31, 2.198, 2.300, migration, 6, 78, 176, 187; scale of 238 2.327 movement, 176; see also 2.77–8 see also 2.304 secondary freshwater fishes, 44 tagging methods, 191, 192 Schistocephalus solidus, 9, 362, 363, selective segregation, 325 see also Salmoniformes 366, 367, 382 self-fertilizing hermaphrodites, 239 Salmonidae (salmonids), 74, 99 schooling behaviour, 8, 269 selfish herd geometry, 288–9 alternative reproductive see also 2.242 semelparous species, 99 strategies, 162–3 Sciaenidae, biogeography, 56, 58 Semicossyphus pulcher (sheephead anadromy, 187 Scleropages, 55 wrasse), 313 biogeography, 49, 58 Scomber scombrus (mackerel), 1, 8, Semotilus atromaculatus (creek egg size, 151 72 chub), 259, 260, 287, 289 hooked jaws (kype), 230 dynamic lift, 72 senescent stage of life cycle, 98, 99 lake species pH sensitivity, 323 migration, 186; scale of Sepia officinalis (cuttlefish), 287 offspring size/numbers, 158 movement, 176 sequence data, systematic parasitism, 376; behaviour see also 2.38, 2.49, 2.176, 2.345 investigations, 18 change, 368; economic impact, Scombridae (mackerels) Serranidae, biogeography, 56, 58, 59, 377; immune defences, 382 carangiform swimming, 75 61, 344 sex determination, 218 piscivory, 269, 270 Serranus baldwini (lantern bass), spawning aggregate swimming adaptations, 77; 236 overexploitation, 240 dynamic lift, 72; tissue lipids, Serranus fasciatus (barred serrano), temperature influence on static lift contribution, 73 236, 237 distribution, 324 Scombroidei, 35 Serrasalmus spp. (piranha), 268, Salmoniformes, 19, 22, 29–30, 328, Scopelomorpha see 269 329 Myctophiformes sewage effluent pollution, 376 salt tolerance of fish taxa, 44 Scorpaenidae (scorpionfishes) sex change, 7, 238–9 Salvelinus, 207 Batesian mimicry, 285 impact on fishing pressure Salvelinus alpinus see Arctic char biogeography, 56, 61 responses, 241 Salvelinus fontinalis (brook trout; catadromy, 188 size-advantage model, 238, 238–9 brook char), 29, 156, 290 Scorpaeniformes, 36 sex chromosome systems, 218 egg size, 161 Scorpaenoidei, biogeography, 58 sex determination juvenile density-dependent scorpionfishes see Scorpaenidae environmental factors, 101–2 mortality, 136 SCUBA studies, 345 genetic basis, 218 life history: fecundity, 154, 155; temperate demersal sex roles, 235 variation, 151–2 environments, 348 sex steroid hormone receptors, 102 recruitment–spawner abundance sculpins see Cottidae sex-specific fisheries, 241 relationship, 126–8, 127 Scyliorhinidae (catsharks), 24 sexual maturation, 6, 97, 99 see also 2.372, 2.377 sea bass see Dicentrarchus labrax ’critical size’ relationship, 97 Salvelinus malma (Dolly Varden sea lamprey (Pteromyzon marinus), see also age at maturity char), 324 158, 176 sexual selection, 225, 230–5 Salvelinus namaycush (lake trout) sea lice, 382 body size, 230–1 functional responses, 273 sea moths (Pegasidae), 35 competition for mates, 230–1 sex determination, 218 sea otter (Enhydra lutris), 313 environmental influences, 227 sand goby (Pomatoschistus sea serpents, 32 fighting weaponry, 230 minutus), 155, 227, 233 sea urchin population explosions, mate choice, 230, 231–5 sand tiger (Odontaspis taurus), 151 313 operational sex ratio, 226 Sarcopterygii, 26 seadragons, 35 ornaments, 231, 233–4 Sarda sarda (bonito), 72 biogeography, 61 parasites as selective agents, 366 Sarotherodon galilaeus (St Peter’s see also Syngnathidae sex roles, 235 fish), 228 seagrasses, 347, 351 sperm competition, 231 Sarotherodon melanotheron seahorses (Hippocampus spp.), 251 see also breeding systems; mate (black-chinned tilapia), 235 monogamous breeding systems, choice Sarpa salpa, 308 237 Seychelles, 60 satellite tagging, 183 see also Syngnathidae shad, anadromy, 176, 187 sawfishes (pristids), 25 search tactics, 271, 291 shark rays (rhynids), 25 scales grouped prey, 288 sharks annual marks/rings (annuli), see also 2.234 denticles, 77 105–6; see also 2.72 seasonal factors migration, 176 ultrastructure, systematic growth effects, 111, 112 sheephead wrasse (Semicossyphus investigations, 17 habitat use changes, 178 pulcher), 313 410 Index sheepshead minnow (Cyprinodon Sparidae, biogeography, 61 Stegastes nigricans (damselfish), variegatus), 216, 368 Sparisoma spp. (parrotfish), 308 233 shelf-sea migratory fish, 181, 182, spawning aggregations, 236 Stegastes partitus (bicolor 183, 185 fisheries damselfish), 233 Shepherd stock-recruitment model, exploitation/overexploitation, stenohaline habit, 78 125 239, 240–1 Stephanoberyciformes, 19, 23, 31, see also 2.178, 2.179 Gadus morhua (Atlantic cod), 7, 33 shoaling, 287 236, 349 Sternoptychidae (hatchetfishes), 30 parasite infection-related spawning areas biogeography, 56 changes, 369 bluefin tuna (Thunnus thynnus), steroidogenic enzymes, sex shrimpfishes (Centriscidae), 35 183 determination, 102 Siluriformes (catfishes), 19, 22, 29 cod (Gadus morhua), 348–9 stichaeids, biogeography, 58 biogeography, 53, 54 eels, 180 sticklebacks see Gasterosteidae egg size, 151 habitat structure, 349 Stizostedion vitreum (walleye), 139, size see body size homing behaviour, 190 270 skates, 25 mate choice, 234 density-dependent mortality see also Rajiformes North Sea plaice, 185 during preadult stages, 136, 138 skeletal tissue, specific gravity, 72 spawning columns, 349 stock skin friction drag, 76 spawning runs, 99 biological definition, 211 skipjack tuna (Katsuwonus speciation, diadromous populations, enhancement programmes: pelamis), 72, 76 193 genetic diversity impact, see also 2.241 species, 18–19 215–17; outbreeding slickheads see Argentiniformes conceptual problems, 19–20, 20 depression, 216; piscivore slimy sculpin (Cottus cognatus), introductions see exotic species introductions for 227 introductions biomanipulation, 279; see also slow (red) muscle fibres, 76, 76 molecular genetic methods of 2.376–7, 2.377, 2.382–3 smallmouth bass (Micropterus identification, 206–7 genetic analysis, 210–15; albacore dolomieui), 154, 241 specific gravity, 72 tuna (Thunnus alalunga), 214; Smegmamorpha, 31 sperm bigeye tuna (Thunnus obesus), smelts see Salmoniformes competition, 225, 231 214, 215; chum salmon snappers (Lutjanidae) morphology, systematic (Oncorhynchus keta), 211–12, biogeography, 344 investigations, 17 212; cod (Gadus morhua), catadromy, 188 Sphyrna (hammerhead shark), 24 212–13; yellowfin tuna snooks, catadromy, 188 spines, 286 (Thunnus albacares), 213–15, snotnose fishes see spiny dogfish (Squalus acanthias), 214 Ateleopodiformes 102, 104, 151 structure determination, 200 social behaviour, 225 spiny eels see Notacanthiformes stock-recruitment models, 6, 124–5 sockeye salmon see Oncorhynchus spiny lobster (Jasus edwardsii), 207 see also 2.185–6, 2.283 nerka spiny lobster (Panulirus stomach, 90 soft flesh disease, 377 marginatus), 215, 313 stomach contents analysis, 115, 278 Solea solea (sole), 36, 348 spot (Leiostomus xanthurus), 273 food web construction, 305–6, migratory behaviour, 186 Squalea, 24 307; advantages/disadvantages, see also 2.18, 2.353 squalene, 25 309 Soleidae, biogeography, 61 static lift contribution, 73 gravimetric analysis, 115 Solenostomidae (ghost pipefishes), Squaliformes, 19, 21, 24, 25 numerical analysis, 115; 35 Squalus acanthias (spiny dogfish), frequency of occurrence, 115; Somniosus microcephalus 102, 104, 151 size of prey, 115 (Greenland shark), 25 Squatiniformes, 19, 22, 24, 25 volumetric analysis, 115 sonar tracking, 185 squid (Loligo vulgaris), 287 see also 2.156, 2.160, 2.185, 2.277 South African marine fauna, 60–1 squirrelfishes (Holocentridae), 33 Stomiidae (dragonfishes), 30 South American freshwater stable isotope studies Stomiiformes, 19, 22, 30 biogeography, 52–4 advantages/disadvantages, 309; stream fishes South American lungfish see isotope routing, 308 density-dependent mortality Lepidosireniformes food webs: advantages, 307–8; during preadult stages, 136, 141 Southern Asia freshwater fingerprints, 306, 307, 308; habitat selection, 258 biogeography, 54 long-term changes, 307; species distribution patterns, 323, southern flounder (Paralichthys see also 2.219, 2.349 323; temperature influence, lethostigma), 273 migration patterns, 306 324 southern pouched lamprey stalking behaviour, 268 striped bass (Morone saxatilis), 76, (Geotria australis), 187, 193 status signals, 231, 233, 234 113, 160, 274, 366 Index 411

recruitment–spawner abundance Syngnathidae, 35 latitudinal gradients in temperate relationship, 129, 129 biogeography, 56, 58, 59, 61 demersal environments, 348 Stromateoidei, 35 Syngnathus typhle (pipefish), 231, meristic characters influences, Strongylocentrotus franciscanus, 235, 236 101 313 Synodontis multipunctatus, 237 operational sex ratio influence, Strongylocentrotus polyacanthus, SypI, 213 227 313 systematics, 4, 5, 15–36 oxygen availability influence, 89 Strongylocentrotus purpuratus, 313 character analysis, 17–18; phenotypic sex determination, sturgeons see Acipenseriformes objectives, 17 101–2, 218 subcarangiform swimming, 75 cladistic analyses, 15, 16–17 predation risk effect, 293 submersibles studies, 345 evolutionary taxonomy, 15, 17 sea bass responses: seasonal temperate demersal phenetic classifications, 15, 16 movement, 184; spawning, 183 environments, 348 species distribution influence, 321 substocks analysis, 5 Tachinoidei, 35 species performance effects, see also stock, genetic analysis tagging methods, 105, 191–2 323–4, 324 substratum structural degradation, molecular, 207 see also 2.93, 2.239–40, 2.241 348 parasites as biological tags, 383 Terapontidae (therapons) suckers see Cypriniformes tagging studies catadromy, 188 summer flounder, ambush age determination, 105 freshwater biogeography, 55 behaviour, 268 blue shark migrations, 189 terrestrial biogeographic realms, 44, Sundasalanx, 28 bluefin tuna migrations, 183, 45 sunfishes (Lepomis spp.), 325 189 territorial behaviour, 164, 234, 236, surfperch (Cymatogaster), 151 homing behaviour, 190 237, 349 surgeonfish (Acanthurus), 91 plaice migrations, 185, 186–7 refuge use in prey, 291–2 swamp eels (Synbranchoidei), 34 sea bass migrations, 183–4 tertiary freshwater fishes, 44 swimbladder see also 2.298, 2.300 testes size, 231 development, 73 tarpon (Megalops atlanticus), 28 tethering experiments, 278 physiology, 73–5, 85; rete tarpons see Elopiformes tetracycline tagging, 191 mirabilis (counter current teeth, 91, 268, 270 see also 2.347 system), 74, 74 Teleostei, 27 Tetraodontidae, biogeography, 61 swimming, 75 adaptive radiation, 71 Tetraodontiformes, 19, 23, 31, 36 adaptations, 71, 75–7; chasing egg size, 151 tetrapods, 19, 26 piscivores, 268 Teleostomi, 21, 26 tetras see Characiformes aerobic metabolism, 77 Telmatherinidae, 34 Thalassoma bifasciatum (bluehead anaerobic metabolism, 77 temperate coastal fisheries, 345 wrasse), 164, 227, 241 burst, 77 temperate demersal environments protogynous hermaphroditism, drag forces, 76–7 latitudinal temperature gradients, 218 foraging capacity relationship, 348 Thalassophryninae, 33 275, 334 recruitment variation, 352 therapons see Terapontidae force generation, 75–6 spawning aggregations, 349 thermal pollution, 376 glide–power cycle, 77 temperate littoral biodiversity, threat signals, 231 impact of parasite infection, 367, 344–5 threespine stickleback 368, 369 demersal species, 348–50 (Gasterosteus aculeatus), 155, migration behaviour, 189; study methods, 345 156, 231, 256, 286, 288, 292, 293 upstream migration, 193 temperate parasite communities, parasitism, 362, 363, 367 prolonged, 77 363–4 thresher sharks (Alopidae), 24–5 respiratory/cardiovascular temperate production, 350–1 Thunnus alalunga (albacore tuna), adaptations, 77, 84 temperature stock structure genetic sustained, 77 cold adaptation see cold analysis, 214 swordfish (Xiphias gladius), 251 adaptation Thunnus albacares (yellowfin tuna) see also 2.21, 2.22, 2.319, 2.320 critical thresholds, 87 stock structure genetic analysis, symbiotic luminous bacteria, 92 development influence: 213–15, 214 Symphodus melops incubation time, 99–100, 100; see also 2.21, 2.194, 2.241 (corkwing wrasse), 163 larvae, 101 Thunnus obesus (bigeye tuna) Symphodus tinca (peacock wrasse), egg size influence, 161 stock structure genetic analysis, 164 food intake influence, 110, 111 214, 215 symplesiomorphy, 16 frontal zone boundaries, 350 see also 2.21, 2.22, 2.23 synaptomorphy, 16, 17 growth influence, 110–11, 111, Thunnus thynnus (bluefin tuna) Synbranchoidei (swamp eels), 34 114; high-latitude populations, migratory behaviour, 182–3, 189 Synbranchyformes, 19, 23, 31, 34 114; seasonal variations, 112 see also 2.21, 2.23, 2.204, 2.330 412 Index

Thymallus thymallus (grayling), biodiversity, 343–4, 346; viruses, 360, 361 166 demersal species, 346–8, 351; viscosity of water, skin friction drag, tidal currents, 181, 182 study methods, 345 76–7 influence on migratory behaviour, recruitment variation, 352 visibility indicator models, 231 180, 181, 182, 186, 189 tropical marine biogeography visual senses tidal stream transport, 181 eastern Pacific, 58–9 foraging capacity relationship, plaice migrations, 185–6, 186 Indo-Pacific, 59–60 275, 334 tiger fish (Hydrocyon vittatus), 268 western Atlantic, 57–8 impact of parasite infection, 367 tiger shark (Galeocerdo cuvier), 24 tropical parasite communities, piscivorous fishes, 270, 275 Tilapia spp., 153 363–4 see also 2.22 tilapias tropical production, 350–1 viviparity, 98 DNA analysis for species troutperches see Percopsiformes growth–maturity–longevity identification, 207 trouts see Salmoniformes (GML) plots, 103 parasite control, 382 true minnows see Cyprinidae von Bertalanffy growth function, 5, tissue RNA as condition indicator, trumpetfish, stalking behaviour, 103–4, 104, 158 109 268 see also 2.111, 2.112, 2.114–15, toadfishes see Batrachoidiformes trunkfishes, 75 2.115, 2.116, 2.190, 2.194, topminnow (Poeciliopsis spp.), trypsin, 90 2.195, 2.196, 2.197, 2.199, asexual species, 237–8 tunas, 251, 267, 342 2.201, 2.205, 2.221, 2.279, 2.280 torpedinids (electric rays), 25 chasing behaviour, 268 vulnerability to fishing, 4–5 Trachichthyiformes, 31 DNA analysis for species compensatory reserve, 124 Trachurus murphyi (Chilean jack identification, 207 life-history characteristics, 6, 7 mackerel), 2 migration, 176, 180; maximum reproductive rate trade-offs see also 2.241–2 calculation: meta-analysis, habitat choice-related predation swimming, 72, 75 129–30; sockeye salmon risk, 258–60 tagging methods, 192 (Oncorhynchus nerka), 126 ontogenetic niche shifts, 327–8, see also 2.21, 2.22, 2.23, 2.49, sex changing reef fishes, 241 329 2.242 refuge use, 290, 290–3 tunnel respirometry, 77 Wallace, 44 transgenic fish introductions, 217 two-spotted goby see Gobiusculus Wallace’s zoogeographic regions, 44, trematode parasites, 360, 361, 368, flavescens 45, 49, 54 369, 377 walleye (Stizostedion vitreum), 139, human pathogens, 377, 378–9 Umbridae (mudminnows), 30 270 Triakidae (hound-sharks), 24 upland bullie (Gobiomorphus density-dependent mortality tribe, taxonomic, 18 breviceps), 233, 369 during preadult stages, 136, Trichopsis vittata (croaking upwelling areas 138 gourami), 231 pelagic species’ habitat, 349–50 water currents, migration impact, triggerfish (Balistidae), 251 phytoplankton primary 179–82 see also 2.306, 2.344 production, 350 ocean currents, 180, 180–1 Trigla spp. (gurnard), 348 productivity, 303–4, 304 tidal currents, 180, 181 Triglidae, 348 Uranoscopidae, biogeography, 61 wax esters, static lift contribution, triglycerides, static lift urea 73 contribution, 73 elasmobranch tissues, 72, 81 Weberian apparatus, 29 trimethylamine oxide, 72, 81 pelagic larvae, 72 weight–length relationships, growth Trimmatom nanus, 35, 151 urine of freshwater fish, 79 assessment, 107–9 triplefins, haemoglobins, 87, 88 Urolophidae, biogeography, 61 Weisberg back calculation method, Tripterygiidae, biogeography, 61 Uvulifer sp., 238 107 trophic cascades, 262, 277–8, 313, western boundary currents, 180 313 vaccines for parasite control, 383 western Palaearctic freshwater see also 2.328, 2.342, 2.344–5 vascular plants, 350 fauna, 49–50 trophic level, 303 vascular system, 84 whale shark (Rhincodon typus), 24, see also 2.206, 2.217–19, 2.286 vendace (Coregonus albula; cisco), 151, 251 trophic morphs, 270–1 335, 336 whalefishes see trophodynamics, 301–2 venomous spines, 33 Stephanoberyciformes fisheries modelling, 315–16 vertebral numbers, 101 whaler shark (Carcharhinus), 24 predation effects, 331–2 Vibrio, 92 whirling disease, 368 trophic transfers within food vicariance, 44–6, 46 white muscle webs, 310, 311, 312 cladistic biogeographic study fast muscle fibres, 76, 76 tropical coastal fisheries, 345 approaches, 47 glycotic enzymes as condition tropical littoral environment Vinciguerria spp. (lightfish), 151 indicators, 109 Index 413

‘white spot’ disease Xiphophorus (live bearers), 53 yolk, 98, 99 (Ichthyophthirius multifilis), Xiphophorus nigrensis yolk-sac larva, 99 361–2 (pygmy swordtail), 163 white stickleback zebra danio, zebrafish (Gasterosteus sp.), 153 Y-chromosome-specific probes, 218 (Brachydanio rerio), 29 whiting see Merlangius merlangus year-class success variability Zeiformes, 19, 23, 31, 33 Wiley, 20 match/mismatch hypothesis, 137 Zoarcidae, biogeography, 58, 62 wind stress, pelagic stage density- see also 2.72, 2.72–3 zooplanktivores, 277–8 independent mortality, 139 yellow flounder (Pleuronectes zooplankton wrasses see Labridae ferruginea), 157 lakes food chains, 329, 333; fish yellow perch see Perca flavescens predation impact, 331, 332; xenobiotics, phenotypic sex yellowfin tuna (Thunnus albacares) population oscillations, 335 influence, 102 stock structure genetic analysis, marine food chains, 304, 350, Xiphias gladius (swordfish), 251 213–15, 214 351 see also 2.21, 2.22, 2.319, 2.320 see also 2.21, 2.194, 2.241 trophic cascade model, 277–8